A word on why I write it. I have started four companies, and I am building my fifth: this one, a media company I run almost entirely solo, on a fleet of about thirty AI agents that draft this newsletter, scan my inboxes, prep my client work, and write my morning brief. I am not reporting on this shift from the sidelines. I am living it, and I share what works and what breaks, openly, as I go.

Subscribe now

There is a strange thing that happens when you ask AI for business advice.

It is almost too helpful.

Tell it, “I spend too much time following up with leads,” and within seconds you have a proposed CRM workflow, three automated sequences, an AI qualification agent, and a list of software you should connect.

It feels productive.

But there is a problem.

Nobody stopped to ask whether follow-up was actually the reason you were losing sales.

Maybe you do not have enough qualified leads. Maybe the leads are good but the offer is wrong. Maybe prospects are interested but there is no clear next step. Maybe your CRM is perfectly capable and nobody uses it consistently.

You complained about follow-up.

AI heard: automate follow-up.

That is the difference between answering a question and diagnosing a problem.

And I think it explains a lot of disappointing AI implementations.

The mechanic who replaces whatever you point at

Imagine taking your car to a mechanic because it shakes at 60 mph.

You say, “I think I need new tires.”

A bad mechanic says, “Absolutely,” and starts replacing them.

A good mechanic asks questions.

When did it start? Does it happen while braking? Is it in the steering wheel or the whole car? Did you hit anything recently?

Then he inspects the car.

Because the customer’s proposed solution is not the same thing as the diagnosis.

We understand this instinctively with doctors, mechanics, accountants, and good consultants.

Yet with AI, we routinely do the opposite.

We tell it:

Help me automate client onboarding.

Or:

Build me an AI sales system.

Or:

I need a better content strategy.

And the machine obliges.

The better the model becomes, the more convincing the answer sounds.

That can actually make the problem worse.

A sophisticated answer to the wrong problem is still the wrong answer.

I realized I was optimizing the prompt instead of the thinking

For a long time, the conventional advice was to write better prompts.

Give AI a role. Provide context. Specify the output. Add examples. Tell it to think carefully.

All useful.

But I became interested in something slightly different:

What if I could change what AI is required to do before it is allowed to give me an answer?

Instead of trying to make the answer better, could I improve the process that produces the answer?

Think about hiring a senior operator.

You would not value her because she can produce longer reports than a junior employee.

You value her because she knows when not to answer yet.

She asks the uncomfortable question.

She notices that the problem you described is probably a symptom.

She wants to know what you have already tried.

She asks what success actually means.

And occasionally she says:

I don’t think that’s the problem we should solve.

That is much closer to how I want AI to work with me.

So I gave AI a process

The framework I ended up with looks roughly like this:

Outcome → Current State → Constraint → Working Hypothesis → Economics → Solution → Measurement

The order matters.

Start with the outcome

“I want to automate sales” is not an outcome.

Do you want more qualified conversations? Higher conversion? Faster follow-up? More revenue? Less founder time?

Those can lead to completely different systems.

Understand what happens today

Before redesigning something, AI should understand what already exists.

What works? Where does the process stop? What requires judgment? What is repetitive? Where are people waiting? Where are they hunting for information?

Otherwise, AI tends to design the business it imagines rather than improve the business you actually have.

Find the constraint

This may be the most important part.

Suppose you have 30 qualified prospects each month but only 12 ever reach a sales conversation.

You could spend months doubling your audience.

Or you could ask why 18 people who were already qualified disappeared.

It is like adding water pressure to a hose with a kink in it.

More water isn’t necessarily the answer.

Find the kink.

Then I added one rule that changed everything

Before recommending a solution, I require AI to state a working hypothesis.

Not a conclusion.

A hypothesis.

For example:

Working hypothesis: The primary revenue constraint is not lead generation. It is inconsistent movement of qualified prospects from initial interest to a sales conversation.

Then it has to answer two more questions:

Why do I think this?

And, more importantly:

What would prove me wrong?

That second question matters.

Without it, AI is remarkably good at constructing an argument for whatever it has already decided.

Humans are pretty good at that too.

Requiring disconfirming evidence changes the conversation from:

“Give me a convincing recommendation.”

to:

“Give me an explanation we can test.”

That is a much higher standard.

Only then do we talk about AI

Here is another rule I built into the system:

Do not assume automation is the solution.

Sometimes the best automation is deleting the task.

Sometimes the process needs to be standardized.

Sometimes the same decision is being made 40 times when it could be made once as a rule.

Sometimes information is scattered across five places.

Sometimes the work simply belongs to someone else.

Only after those possibilities have been considered should we ask what AI can do.

My preferred sequence is closer to:

Eliminate → Simplify → Standardize → Clarify ownership → Automate → Add AI → Build custom technology

It is the business equivalent of not buying a robotic lawn mower before checking whether you actually need to mow that part of the yard.

And building it isn’t enough

This is another place where I think businesses will get AI wrong.

An automation runs successfully.

Everyone celebrates.

But “it ran” tells you almost nothing.

Imagine hiring a salesperson and measuring performance by the number of emails they sent.

That is activity, not value.

The same distinction applies to AI.

I now think about four different states:

BUILT → RUNNING → WORKING → CREATING VALUE

Those are not synonyms.

An AI follow-up system might be built and running perfectly while annoying good prospects and reducing conversion.

A client-onboarding agent might process every client while creating more work for the team fixing its mistakes.

So before building, the system should define what working means.

Maybe follow-up falls from 48 hours to 8.

Maybe founder administration falls from six hours a week to two.

Maybe qualified-lead conversion moves from 12% to 17%.

Now we have something to test.

This is becoming my default way of working with AI

I don’t want AI to agree with me faster.

I want it to make it harder for me to solve the wrong problem.

I want it to challenge assumptions, distinguish facts from hypotheses, look at economics, identify what should remain human, consider how a system could fail, and tell me what evidence would change its recommendation.

In other words, I don’t just want a better answer.

I want a better thinking process before the answer.

So I turned all of this into a reusable set of AI instructions.

It forces the model to begin with three high-value questions, diagnose before recommending, state a falsifiable working hypothesis, test the economics, choose the simplest intervention, separate human judgment from automation, define failure conditions, and establish how we will know whether the result actually worked.

That’s the what.

For paid subscribers, the rest of this article is the how.

🔒 PAID: Install the AI Thinking System

Below is the complete instruction set I use, ready to copy into your AI.

I’ll also show you:

• how to install it

• how to bypass the three questions when you want an immediate answer

• how to make AI interrogate the problem further when necessary

• how to use it to diagnose a real business problem

• and the complete master instruction set

The goal isn’t to make AI give you more answers.

It’s to make it earn the right to give you one.

Subscribe now

🔒 PAID: Install the AI Thinking System

If you made it this far, the useful part is not just understanding the framework.

It is making your AI behave this way by default.

The goal is to create a reasoning layer that sits between your question and the answer.

Instead of:

Prompt → Answer

you get:

Prompt → Clarify → Diagnose → Form hypothesis → Test economics → Recommend → Measure

That changes the interaction completely.

Read more

A word on why I write it. I have started four companies, and I am building my fifth: this one, a media company I run almost entirely solo, on a fleet of about thirty AI agents that draft this newsletter, scan my inboxes, prep my client work, and write my morning brief. I am not reporting on this shift from the sidelines. I am living it, and I share what works and what breaks, openly, as I go.

Contents

Why Advice Fails and Systems Win

What this Build Sprint Series is

What Consultants and Operators should expect

Mini Case Study: The Solo Consultant Communication Crisis

• 2021 vs 2026 Consulting Approach

• The Machine that I build

• AI Impact

How the Build Sprints Work

Why Advice Fails and Systems Win

Most entrepreneurs suffer from a failure of the human operating system. You finish a high-level coaching program feeling energized, only to watch that momentum dissolve into systemic entropy within weeks. The talented hairdresser understands cash flow perfectly yet remains too busy cutting hair to build a forecast. The baker knows she needs marketing but spends sixty hours a week at the oven. You do not lack knowledge. You lack the functional capacity to execute that knowledge while surviving the daily grind.

For 25 years, I’ve launched, developed, grown, failed, and sold businesses across Finance, Real Estate, Hospitality, and Business Education. The most important and painful lesson I learned is this: building systems that automate repetitive tasks is everything. Knowledge without infrastructure is just noise.

In 2021, I founded a business accelerator program designed to help solopreneurs and small business operators—hairdressers, plumbers, consultants, photographers, bakers, service providers—break through the ceiling trapping them. We advised over 100 businesses that year.

Teaching my first Business Accelerator Cohort. 2021

The pattern was identical across every single one: exceptional at their craft but drowning in everything else. These weren’t people who lacked knowledge or ambition. They lacked the systems to stabilize and grow. A consultant knew exactly what to prioritize but couldn’t find the bandwidth to do it. A photographer had great ideas but no system to execute them. A salesperson could close deals but had no follow-up infrastructure. Everyone was building the plane while flying it.

We diagnosed problems, designed solutions, handed over frameworks. The advice was sound. Execution? 60% at best. By week six, life got busy, frameworks got shelved.

The lesson stuck: Advice without implementation infrastructure doesn’t stick.

I stepped back. Built, grew, and exited a business for $35M. Took a year to ask the real question: what if I didn’t have to accept that 60%? What if infrastructure could be built, not just recommended?

In October 2025, I started working with some of the same operators from 2021. This time, I brought AI. The impact was immediate and profound. The problems were identical: time starvation, coordination chaos, growth ceiling. But the solutions were fundamentally different. Instead of frameworks that required discipline and personal bandwidth, we built machines that executed independent of personal effort.

The operators stayed time-strapped. But now they controlled systems that gave them time back.

This is the shift: AI has moved us from “learning what to do” to “building machines that do it for you.” Strategy is no longer something you learn and implement. It’s hard-coded into your business infrastructure. The goal isn’t to become a better manager. It’s to build an independent system that functions regardless of your personal discipline.

What This Build Sprints Series Is

The cases that follow are actual engagements from either 2021 or 2026. When an engagement happened in 2021, the 2026 approach represents what that same problem would look like solved today with AI infrastructure. When an engagement is from 2026, the 2021 approach shows what the old methodology would have delivered. This illustrates the real gap between advice and infrastructure.

If you’re a consultant: AI has fundamentally changed what you can deliver. It’s no longer an hourly charge. It’s outcome-based. Execution time has shrunk dramatically, both for you and your clients. They now expect a machine built that they can take and drive. That changes your pricing, your accountability, and your competitive advantage.

If you’re a business operator: Here are real ideas from operators like you, using AI infrastructure that actually works. These solve business problems that are incredibly common: time loss, cash recovery, customer consistency, demand generation. These aren’t theoretical. They’re from the last eight months of actual work.

Case Study:

Meet Devon: The Solo Consultant Communication Crisis

Devon runs a solo strategy and operations consulting practice. She’s sharp, credentialed, and her clients love her work.

But she’s stuck. She spends about 3 hours a day roughly 15 hours a week across email, Slack, calendar, and CRM managing communication instead of doing client work, sales, or business development.

The real problem: Every new client adds more communication burden. She can’t scale her practice without drowning in coordination overhead. The constant switching, prioritizing, remembering, and following up has become the ceiling on her growth.

2021: The Advice-Based Approach

The solution focused on helping Devon manage communication more efficiently:

• Email batching

• Response templates

• CRM discipline

• Calendar blocking

• Follow-up routines

• Weekly prioritization

The system worked. Communication workload dropped from 15 hours a week to 11 hours. Devon recovered roughly 4 hours a week.

But Devon still had to operate the system herself. She had to keep the CRM current, maintain the routines, remember follow-ups, and decide what needed attention.

The system improved how Devon worked. It still depended on her to keep it running.

2026: The AI System-Based Approach

Same problem. Completely different solution.

An AI Communication System was developed, built around the tools Devon already uses: Gmail + Slack + Calendar + HubSpot integrated into a lightweight AI infrastructure powered by Claude Cowork.

The architecture:

A Chief of Staff Agent handles the coordination work: organizing communication, tracking commitments, preparing context, drafting routine responses, keeping important information current.

Instead of opening multiple tools and figuring out where to start, Devon begins with a prioritized view of what needs a decision, response, meeting preparation, or follow-up.

How Devon operates it:

Each morning, a priority brief surfaces the 3-5 things requiring her attention, ranked by impact. Client status. Pipeline updates. Meeting prep. Everything else is organized and tracked. Devon spends 30 minutes reviewing context and making decisions. The system runs the rest.

Communication workload drops from 15 hours a week to 8 hours. Devon recovers roughly 7 hours a week.

The system does not replace Devon. It removes more of the coordination work around her, plus helps her prioritize her actions to optimize the time she dedicate to achieve her goals.

The Shift: Before and After

All communication flows through one system. Devon is decision-maker only. System manages the coordination. Time: 15 hours/week managing communication → 8 hours/week. 7 hours recovered.

The Time & Money Impact

Devon charges $175/hour.

Time Savings:

• Pre-AI: 15 hours/week in communication management

• Post-AI: 8 hours/week in communication management

• Time recovered: 7 hours/week

Annual Value:

• 7 hours/week × 52 weeks = 364 hours/year

• 364 hours × $175/hour = $63,700/year in reclaimed billable capacity

Potential: If recovered time gets deployed for business development, add $38,000–$63,000+ in new revenue annually. Most operators recover time but don’t have a plan to deploy it strategically.

How This Works: The Build Sprint Methodology

Every engagement follows a five-stage process called the Build Sprint. It’s simple, repeatable, and designed to move from diagnosis to a running system in weeks, not months.

Phase 1: Design + Define + Diagnose

Diagnose. Real listening. We map the actual workflow—not the one you think you have—and trace visible problems back to their root. A solo consultant spending 15 hours a week in email isn’t the problem. The problem is scattered communication across five tools with no priority system. Visible problems are usually symptoms of invisible system failures.

Define. Once we understand what’s broken, we agree on what success looks like. Measurable success. How many hours per week should this reclaim? What’s the financial impact? What would change in your business if this problem disappeared?

Design. We design the system. What agents do we need? Which tools integrate? How does data flow? This is architecture work. The design phase produces a blueprint everyone agrees on before a single line of code is written.

If you decide to Build it, you will get all spec, the design for you to install and drive the machine.

If you decide to build it together, we work on a new engagement for another $999.

If you decide to have me build it including training and setup, we will work on a proposal.

Phase 2: Delivery + Drive

Deliver. We build the system, test it, install it, and train you to operate it. This is execution. The system gets built, not handed off as a framework to implement. You don’t have to figure out how to use it. You don’t have to maintain it yourself. You open it, it works, you make decisions on what it surfaces. This phase typically takes two to four weeks depending on complexity. When it’s done, you have a running machine, not a homework assignment.

Drive. You operate the system forward. This is the part that’s different from 2021. You’re not struggling to implement something new. You’re running something that already works. The system adapts as your business grows. New tools integrate when you need them. Rules evolve based on what you learn. The operator owns the outcomes, and the system runs the coordination work.

This five-stage approach works because it separates the thinking from the doing. It separates the strategy from the infrastructure. It separates what you have to do from what the machine does for you.

Let’s Connect

Schedule a 15-minute conversation. Listen to what’s keeping you up at night, then get one idea, tool, or workflow you can implement this week.

If it can’t give you a quick solution in 15 minutes, I will get back to you with the answer.

The only ask: If it works, let’s continue the conversation.

Let's connect

Read my Hotel Turnaround Story. At the New Haven Village Suites. Former Residence Inn. 2015

Visit: juansalasromer.com for more information about my programs and assessments.

A word on why I write it. I have started four companies, and I am building my fifth: this one, a media company I run almost entirely solo, on a fleet of about thirty AI agents that draft this newsletter, scan my inboxes, prep my client work, and write my morning brief. I am not reporting on this shift from the sidelines. I am living it, and I share what works and what breaks, openly, as I go.

If you want the right AI solution for your business, not just another list? Book a 15-minute call. I will assess your bottleneck and point you to the right system or app that fits. If it needs more digging, I will come back with the answer on a follow-up call.

Let's connect

Editorial

For most of this year I have not been using AI. I have been managing it. There is a real difference, and it took me a while to see it. A tool is something you pick up when you need it. What I run now is closer to a small staff. Something scans the market before I wake up. Something drafts from my own notes. Something checks the numbers and flags what moved. I sit at the top of it, deciding and approving, the way you would with people. I wrote about the first version of this in how I build my AI chief of staff. This issue is the org chart underneath it, and how you build your own, even if the whole company is just you.

Here is the shape, because the shape is the whole idea.

Three findings from the last eight weeks changed how I would build this, and two of them cut against what is being sold. Competence at individual jobs does not add up to a working system. The agent-to-agent coordination layer everyone is selling has the thinnest evidence behind it of any part of the stack. And in the sectors AI touches most, people have stopped starting companies that intend to hire.

Read together they say something specific. Do not build a swarm. Build one coordinator that routes to you, keep every seat short and testable, and accept that you are the part that cannot be automated. Pick the one seat your business needs most this week, and fill only that one.

Juan

26.8 percent, against 6.4 percent down. In sectors most exposed to AI, solo-type business applications rose 26.8 percent from Q1 2024 to Q1 2026 while a construction-and-wholesale comparison group sat flat at negative 0.4 percent. Over the same window, applications in those AI-exposed sectors carrying markers of intent to hire fell 6.4 percent. The growth came entirely from filings with no employer intent. (Mercatus Center, 17 July 2026, on Census Business Formation Statistics and CPS microdata)

Six clusters, twenty-seven papers. An Oxford synthesis across nineteen benchmarks found that failure compounds non-linearly with task length, that sub-skill competence does not reliably compose into end-to-end success, and that adding scaffolding does not uniformly improve reliability. (University of Oxford, July 2026 preprint, not peer reviewed)

Seven of ten. Recent multi-agent coordination architectures whose headline improvement is smaller than the measured run-to-run noise of the same model on the same benchmark. The clean paired contrast came to five percentage points, with a confidence interval crossing zero. (UT Austin and Uber, 15 June 2026 preprint, not peer reviewed)

Zero. How many of the four confident claims my own system handed me last month survived a check. All four traced to real files. All four were wrong. Nothing was invented, which is the harder failure to catch, and it is why an evaluation step is infrastructure and not a nicety.

79 percent. The share of multi-agent failures blamed on specification and coordination rather than the models, quoted all summer as this year’s finding. It comes from a paper published in March 2025. The work is solid and I would have used it, except seventeen months is not recent, and a new date on an old number is not a new number.

Story 1: Your Agents Can Each Pass and the Company Can Still Fail

What is happening. Researchers at Oxford pulled twenty-seven benchmark, taxonomy and audit papers spanning nineteen benchmarks into six failure clusters (July 2026 preprint). Three conclusions matter more than the taxonomy. Failure compounds non-linearly as tasks get longer. Competence at the sub-skills does not reliably compose into end-to-end success. Adding more scaffolding does not uniformly improve reliability. They are careful about where agents have genuinely improved: single-turn tool selection, short web navigation, narrowly scoped coding.

What it means for you. Notice the shape. Agents are good at short, bounded jobs and degrade as the job gets longer and more connected. That is not an argument against an agent organization. It is an argument for how to draw one: short seats with clear edges, not one agent asked to run a whole function. It also kills the common fix. When your setup is flaky the instinct is another layer, another tool, another wrapper. The synthesis says that does not dependably help. Smaller seats do.

The uncomfortable part if you are buying: each piece can demo perfectly and the assembled thing can still fail, because composition is where it breaks. A vendor demo is a sub-skill test.

What I would do this week. Draw the chart above for your own business. Then ask one question of every box: can this be finished in a few steps with a clear end? If not, split it. Do not automate anything yet. Getting the boxes the right size is most of the work.

Story 2: The Coordination Layer Is the Least Proven Part of the Stack

What is happening. Two researchers, one at UT Austin and one at Uber, asked what the multi-agent field had skipped: before you claim your coordination architecture beat another, how much do two identical setups differ by chance? They measured that floor on one model against a retail benchmark, then held the published claims against it. Seven of ten recent coordination architectures report gains below that floor (15 June 2026 preprint). Their own clean contrast came to five percentage points with a confidence interval crossing zero, which means not significant.

What it means for you. Be careful what this says. It does not say coordination is worthless. It says the evidence that elaborate agent-to-agent schemes work is thinner than the marketing, and some of it may be noise. The practical read: the least proven thing you could build right now is agents negotiating with each other, and it is what is being sold hardest.

Which is why the orchestrator in the chart is drawn the way it is. Its job is not to let agents talk. Its job is to take what came back, decide what needs you, and put it in front of you. The value sits in the escalation, and the escalation ends at a human. I have watched my own system hand me four confident wrong answers in a month. No amount of agents conferring would have caught that. I caught it.

What I would do this week. Build one coordinating routine with exactly one job: gather what happened, sort by what needs your decision, hand you a short brief. Do not let anything act on another agent’s output without you in between. Not yet.

Story 3: AI Is Changing What Kind of Company People Start

What is happening. A Mercatus study published 17 July 2026 compared business formation in sectors most exposed to generative AI against construction and wholesale as a control. From Q1 2024 to Q1 2026, solo-type applications in the exposed group rose 26.8 percent while the comparison group sat flat at negative 0.4 percent. Then the finding that reframes it: applications carrying markers of intent to hire fell 6.4 percent in those same sectors, even as total applications rose. Census Business Formation Statistics and Current Population Survey microdata agree.

What it means for you. The story is not that more people are going solo. It is that where AI reaches, the kind of company people choose to start has changed. Fewer are built to hire. More are built to stay one person. That lands on you directly: if your growth plan never included headcount, your org chart is not a metaphor. It is the staffing plan, and the seats get filled with agents or they stay empty.

That is also the honest limit. This is early evidence of a correlation in exposed sectors, not proof AI caused it, and the authors say so. What it does establish is that you are not unusual for planning a company that runs without employees.

What I would do this week. Take your chart and pick the single function that eats the most of your week and needs your judgment the least. Recurring admin, first-draft content, lead follow-up. That is the first seat. One. The point is not a full company by Friday. It is proof that one box can run without you.

Move 1: Draw the chart, then size the boxes. Your functions, an orchestrator, you at the top. Split any box that cannot be finished in a few clear steps. Long connected jobs are where agents fail, so the sizing is the work.

Move 2: Fill the orchestrator seat, and point it at yourself. One routine that gathers, sorts by what needs your decision, and briefs you. Not agents conferring. The escalation to a human is the part with evidence behind it.

Move 3: Stand up one specialist, the highest-drag lowest-judgment one. Give it a job, the access it needs and no more, a place its output lands, and an off switch. If you cannot write those four things down, it is not ready.

Move 4: Put a test on it that can fail. Not “does this look right,” which the system will always say yes to. A number that clears a bar or does not. Resist adding a wrapper when something is flaky. Check what stays in your hands against the six things that should stay human.

None of these needs a developer, and none starts with buying an agent.

PROMPTS

The Agent Org Chart. Walks away with their business drawn as an org chart of functions, an orchestrator, and a human accountability layer, every box sized against the short-and-testable rule and the safest one marked to fill first.

GO TO PROMPT

The Orchestrator Brief. Walks away with the spec for one coordinating routine: what it gathers, how it sorts by what needs a decision, and what it escalates to a human.

GO TO PROMPT

The Agent Job Description. Walks away with one specialist defined before it is built: job, access, output, off switch, and the test that has to pass. If they cannot fill it in, the agent is not ready, and that is the useful answer.

GO TO PROMPT

Run in order they produce three documents: a map, a spec, and one job description. This is the thinking that happens before anything gets built, and it is the part most people skip on their way to a mess.

If you run a business and want to know which seat is costing you most, the free Diagnostic takes fifteen minutes and names your biggest constraint and what it costs. learn.buildtothrive.co/diagnostic. When you want the org chart mapped rather than guessed, Find the Operational Drag is the $999 assessment that walks your operation function by function and tells you which seat to fill in what order.

If you would rather build it with someone, take fifteen minutes with me and we will draw your agent org chart together and pick the first seat.

Let's connect

Check out my Builder’s Store (Work in progress)

…and remember

Build assets. Create freedom. Thrive on your terms.

Juan

Oh…I almost forgot.

Want more free resources? Fifty operators shared the prompts behind the first six months of our AIBlueprint Edition. 50 Creators share their rockstar prompts →

Liked, downloaded and shared by more than 1000 people…you can’t go wrong!

Subscribe now

A word on why I write it. I have started four companies, and I am building my fifth: this one, a media company I run almost entirely solo, on a fleet of about thirty AI agents that draft this newsletter, scan my inboxes, prep my client work, and write my morning brief. I am not reporting on this shift from the sidelines. I am living it, and I share what works and what breaks, openly, as I go.

TLDR

A recurring bill is easy to see and easy to argue about. What the bill is buying is usually invisible, and the expensive part is almost never the amount. It is whether you could leave. The AI make or buy decision.

Last summer I went through my own AI spend line by line, expecting to find one expensive tool I could cut. That is not what I found. The cost was spread across a dozen small automatic things that were each too minor to question on their own. I cut the bill roughly in half without changing any of the work.

What stayed with me was not the savings. It was that I had been paying that bill every month for months and could not have told you what most of it was for. Not because I am careless with money. Because a subscription is designed to stop being a decision.

I thought about that this week reading a thread I have not been able to stop turning over.

The thread

A man in a small town runs a mobile auto-detailing business. He has been paying $379 a month for six years for what he was told is a custom booking system. His nephew opened the developer tools on the site.

The custom booking system is a single script that embeds Square. The service-area map is a screenshot. There is no custom backend at all.

The provider has now quoted him $2,400 for a platform upgrade, after which the monthly goes to $469. Changing the price of a wash package costs $95.

The internet did what the internet does. Ninety-three upvotes and a hundred and sixteen comments, most of them arguing about whether $379 is too much.

What most of the room got right, and why it still misses

The obvious answer is that he is being overcharged, and plenty of people said so. One version of it sat at eighty-two upvotes.

Then something more interesting happened. The top comment of the entire thread, at two hundred and forty-six upvotes, told the nephew not to touch it.

“If you don’t know the answer to this question you are not qualified to be managing it. I would charge a bit more than your current guy. Not messing with it is the least risky option.”

Another, at forty-four: “OP is out of his depth here. So fixated on a few hundred dollars a month he is about to wipe out his uncle’s business.”

Someone pointed out the six years of accumulated search history a rebuild would throw away. Someone else called the nephew “the one who can do it cheaper that the prophecies foretold.”

I want to be careful here, because that camp is right about something real. The instinct to rebuild the thing yourself, now that the tools are cheap, is the second most expensive mistake available in that situation. It is not the first.

The three comments nobody upvoted

Buried under all of it, at single-digit upvotes, three people were having a completely different conversation.

“The 379 a month isn’t really the problem, and neither is the 2400. What he’s paying for is not being able to leave.”

“The biggest issue here is not what your uncle is paying this vendor but that he doesn’t have the keys to his business.”

“The thing I’d check before touching anything is whose name the domain and the business email are actually in. Six years with one provider usually means both sit inside their account, and that is the real leverage, not the booking widget.”

Three people, arriving from three directions, at the same answer. The number is not the problem. The lock-in is the problem, and the number is what the lock-in costs.

None of them wrote down what to check.

The broader pattern

I think this shows up in a lot of operations, and almost never as a website.

It looks like the bookkeeper who has been doing your books for nine years and is the only one who understands your chart of accounts. The scheduling tool your front desk built a workaround inside of. The one spreadsheet that runs your pricing, which lives in an employee’s personal drive. The email list you cannot export in a usable form.

In each case there is a visible number and an invisible dependency, and the operator is looking at the number because the number arrives once a month with a subject line.

Here is the test I have started using on my own business. It is not about cost.

If this provider stopped answering tomorrow, what would I lose that I could not replace in a week?

If the answer is nothing, the bill is just a bill and you can negotiate it like one. If the answer is my customer list, my domain, my search history, or my ability to take a booking, then you are not buying a service. You are renting the ability to keep operating, and the price of that is whatever they decide next year.

Why this matters now

Two things changed at once and they pull in opposite directions.

The cost of rebuilding almost anything fell through the floor. That is real, and it is why the nephew’s instinct is not stupid.

But the cost of a bad migration did not fall at all. Six years of search history, a customer database you assumed was yours, a business email address printed on the side of a van. AI made the build cheap. It did nothing to make leaving safe.

Which means the gap between those two things is now the most expensive place in a small operation, and almost nobody is measuring it.

So should you just rebuild it with AI?

Everyone reading this has already had the thought, so let me take it seriously rather than waving at it.

A weekend and a decent model will get you a booking page. That was not true three years ago and it is true now. The people saying otherwise are defending something.

But look at what the rebuild accomplishes if nothing else changes. The uncle stops paying a provider he could have fired, and starts depending on a nephew he cannot. The lock-in did not go away. It moved, and it moved to the person hardest to replace. That is what the do-not-touch camp is sensing when they say he is not qualified. They are right about the risk and wrong about the reason.

And the same is true of anyone else he brings in, including someone he pays properly. The question is never who builds it. It is whose name is on it when they are finished. Set up inside the uncle’s own accounts, a nephew has done him a favor. Set up inside the nephew’s, he has quietly become the new provider at a better price and worse terms, because now the exit is a family conversation.

This is the make-or-buy call, and the reason it goes wrong is that most operators run it on the whole vendor at once, when the two halves of what they are buying have opposite answers.

The inventory sorts it, because it separates two kinds of line that look identical on an invoice.

Records are the domain, the business email, the merchant account, the customer list, the analytics history. A record is never a make-or-buy decision. You do not build one and you do not really buy one. You get your name on it. That is paperwork, and no model has ever made paperwork happen.

Surfaces are the site, the booking flow, the forms, the page that shows the service area. Those are cheap to rebuild now, genuinely, and getting cheaper. Surfaces are where make-or-buy is a real question.

Once you know it is a surface, three questions decide it

How often does it change? The man in the thread pays $95 to change the price of a wash package. If he does that twice a year it is an irritation. If he does it monthly it is the entire argument. Frequency is what turns a small fee into a structural cost, and it is the question people forget to ask because each individual fee looks survivable.

What breaks if it is wrong? A page listing your service area can be wrong for a day and nothing happens. A booking flow that takes payment cannot. The closer a surface sits to money, the higher the bar for building it yourself, and the more of that monthly bill was quietly buying you insurance.

Can you describe it completely? This is the one people skip and it is the one that decides. AI builds what you specify. If you cannot write down every rule your booking flow follows, including the ones you have never said out loud, you are not ready to make it. You will also not be able to tell whether anyone else built it correctly.

Two yeses and a low blast radius, build it. Anything else, keep buying and negotiate better.

And the sequence matters more than the answer

Own the records first. Then run make-or-buy on the surfaces, at your own pace, from a position where leaving is possible. Rebuild before your name is on the registrar and you have put a new house on land somebody else owns.

And if you bring somebody in to help, that is the thing to settle before they start. Not the price. Whose name goes on what, and what you are left holding when the work is done. Anyone good will find that an easy conversation. The reaction to being asked tells you most of what you need to know.

One more thing the invoice hides. Part of what you pay a provider is not software at all. It is somebody answering on a Saturday when the booking form breaks. If you build, that person is you. Price it before you decide, not after.

What I would do first

Not renegotiate. Not rebuild. Find out what you own.

Before you touch the invoice, sit down for one afternoon and answer a single question for each piece of your operation: whose name is on it?

Not who manages it. Not who set it up. Whose name is on the account.

For the man in the thread, the answer to one of those lines is probably worth more than the entire six years of $379 payments. And here is the part that should be encouraging rather than alarming: two commenters pointed out that his booking history and customer data sit inside Square, not with the developer. His switching cost is almost certainly far lower than it feels. He does not know that, so he is negotiating from a position of fear he may not actually be in.

That is the whole thing. You cannot negotiate a dependency you have not located, and you cannot price a risk you have not named.

The Keys and Accounts Inventory

One page. One afternoon. Whose name is on each line.

This is the procedure the thread never wrote down. It is yours, no email required.

Work through each row. For each one you need three answers: who is named on the account, who can add or remove that name, and what happens to the business if access stops tomorrow.

Here is a fillable sheet you can print and use The Keys and Account Inventory

How to score it

Mark each line OWNED, SHARED, or THEIRS.

You are not trying to get to all-owned. Plenty of THEIRS lines are fine and normal. You are trying to find the ones where THEIRS and business-stops-tomorrow are true at the same time. That intersection is usually two or three lines, and it is the entire negotiation.

The one move after the inventory

Take your THEIRS-and-critical lines and ask for one thing per line: to be added as an owner, not to take it over.

It is a small ask, it is hard to refuse without saying something revealing, and the answer tells you what kind of provider you have. A good one adds you the same week. The response is the diagnostic.

That is the whole free method. If you only do the inventory and nothing else, you are already ahead of the hundred and sixteen people arguing about the invoice.

SOFT NEXT STEP

I am still running this inside my own business, and the lines that came back THEIRS surprised me more than the ones that came back OWNED. I am sharing the parts that break because those are usually the parts other operators need help naming.

Two things tend to happen after someone fills the sheet in.

The first is that one line turns out to matter more than the other eleven, they sort it out themselves, and that is the end of it. That is the sheet doing its job. You do not owe me anything for it.

The second is that the sheet raises a question it cannot answer. You now know what you could stop paying for. You still have to work out which of those you could actually replace, in what order, and whether any of it is worth a month of your attention. That depends on how your week really runs, which parts only work because you are the one running them, and what your own time is worth against the invoice. A checklist cannot see any of that.

If that is where you have landed, let’s connect over a virtual coffee.

Here is exactly what happens in it, so you know before you book anything.

I ask you three questions. What is taking time away from you. Where could you be making more. Where are your customers being served worse than you would like. You pick the one that matters most right now, and we spend the rest of the call on that one rather than surveying all three.

Then, before we hang up, I come back to you with something. A tool, a prompt, a different way to run the thing. Not notes to action later. Something you can use that afternoon.

Sometimes that solves it and there is nothing else to discuss, which is a good outcome and a common one. Sometimes it turns out to be bigger than fifteen minutes, and then we can talk about going deeper with an AI Leverage Assessment. That is your call, at your pace, not mine.

If it is useful, take fifteen minutes with me and we will look at your week together.

Most of the time we find five hours or more sitting somewhere fairly obvious. Sometimes we do not, and I would rather say so than stretch for it.

I am doing these because seeing how other people actually run their week is how I learn what is really breaking out there. If you get something out of it, a sentence I can quote back is more helpful to me than you would guess.

---

Subscribe now

The Vendor Conversation Pack (Paid subscribers)

Six prompts for the part that happens after the inventory.

The inventory is the easy half. It is a list, and lists are calm.

The hard half starts the moment you have to open the conversation. You are asking someone you depend on to give up a piece of their leverage, usually while you still need them, often while you are not certain what you are entitled to. Most operators either send nothing, or send something that reads like an accusation and gets a defensive reply that costs them six months.

These six prompts cover that arc end to end. They assume you have done the inventory and know which two or three lines are THEIRS-and-critical.

Prompt 1: The ownership ask

Read more

I have spent the last week reading the research on where AI actually breaks, rather than where people say it breaks. Three studies, measuring the same thing from different angles, and they agree in a way I did not expect. The failure is not capability. Models are not getting worse at the work. What collapses is consistency, and it collapses in the exact conditions we all work in: long conversations, a lot of context, and a session that has to remember what we agreed twenty minutes ago. I wrote in June about why the tools kept making me busier, and last month about what happens when you stop collecting tools, both pieces This issue hands you three fixes, and none of them is a better prompt.

If your AI output feels inconsistent rather than bad, this issue gives you the one page that makes it steady. If you deliver work anyone pays for, it gives you the sentence to have ready before a client asks whether AI touched it. And if your sessions have been getting longer because everything needs to stay in one place, it gives you the two-minute habit that replaces them.

Leak One: reliability, not ability. Aptitude barely moves as a conversation runs on. Consistency falls off a cliff, and once a session takes a wrong turn it does not recover. The fix is a page, not a prompt, and it takes an hour once.

Leak Two: the window on the label is not the window you can trust. Accuracy drops well before the advertised limit, and whatever sits in the middle of a long brief is the first thing lost. The fix is where you put things, and it costs nothing.

Leak Three: your buyer already suspects. Seven in ten say they would not trust a report made with AI, and the ones who have used the tools are the more sceptical group. The fix is one sentence, written before the question arrives rather than during it.

All three run on the same idea. Nothing inside a conversation is authoritative, so nothing in there can correct a wrong turn once it happens. Reliability has to come from outside the chat, from something you wrote down while you were thinking clearly. It is not a property of the model. It is a property of what the model can check itself against.

Do one of them this week. What you are after is not smarter answers. It is the same answer twice, which is the thing you have actually been missing.

Juan

16 and 112. In over two hundred thousand simulated multi-turn conversations, leading models lost 16 percent of their aptitude and gained 112 percent more unreliability, an average 39 percent drop across six tasks. Capability held. Consistency did not. (Microsoft Research and Salesforce, ICLR 2026)

More than 30 percent. The performance drop when the thing that matters moves from the start or end of your input into the middle of it. Models have primacy and recency bias, exactly like people. (Stanford, TACL 2024)

30 to 50 percent. Accuracy loss across eighteen frontier models measured well before the stated context limit. A 200,000-token window can be unreliable at 50,000. (Chroma, named lab, not peer reviewed)

70 percent. Buyers who say they would not trust a consultancy report prepared using AI, from a survey of 3,887 companies. Among those who have actually used consultants’ AI tools, 77 percent call it a bubble, against 55 percent who have not.

20 to 30 percent. The widely quoted share of the workday knowledge workers supposedly lose searching for information. I went looking for the source and found four different figures from four vendors and one analysis calling it a myth. Directionally true, badly sourced, and I am not using it.

What is happening. Researchers at Microsoft and Salesforce took standard single-turn benchmarks and converted them into multi-turn conversations, which is how people actually work. Every leading model got worse. The average drop was 39 percent across six tasks.

Then they decomposed it, and that is the finding. Aptitude fell 16 percent. Unreliability rose 112 percent. Their conclusion: when a model takes a wrong turn early in a conversation, it does not recover, because nothing in a chat is authoritative enough to correct it.

What it means for you. If your AI output feels inconsistent rather than bad, that is measured behaviour and not your imagination. It also means the common fix is the wrong one, because sharper prompts address ability and ability is not what broke.

Sit with their explanation, because the fix is inside it. Nothing in a conversation is authoritative. Every statement in a chat carries the same weight as every other statement, including the wrong turn taken at minute four. There is nothing for the model to check itself against, so a small early error compounds instead of getting corrected.

Which tells you exactly where the fix has to live. Not in the conversation. Outside it, in something written down when you were thinking clearly, that does not drift as the session gets long. Reliability is not a property of the model. It is a property of what the model can check against.

What I would do this week. Open one document and write four things: what you sell and to whom, how you decide pricing, what makes a client a good fit, and three decisions you have already made and do not want reopened. Paste it at the top of your next session. That document is the authority the conversation does not have. One hour, and it is the highest-return hour in this issue.

What is happening. Two separate findings, same direction. Stanford measured a U-shaped curve: models use what sits at the beginning and end of what you give them and lose what sits in the middle, with more than 30 percent degradation as the important thing moves inward. And the lab Chroma tested eighteen current models and found accuracy dropping 30 to 50 percent well before the advertised context limit.

What it means for you. Two habits are quietly costing you. Running one enormous session all day instead of shorter ones with the context written down. And burying the constraint that matters in the middle of a long brief, where it is the first thing to fall out. The number on the box is a capacity figure, not a reliability figure.

What I would do this week. Put the things that must always be true at the top of the conversation, not in the middle of paragraph nine. And when a session gets long, stop and ask it to write down what has been decided so far, then start fresh with that summary. Two minutes, and it is the difference between a session that holds and one that drifts.

What is happening. Source Global Research surveyed 3,887 companies. Seventy percent said they would not trust a consultancy report prepared using AI. Almost a third said learning a firm had used AI would shake their confidence in that firm.

The part that inverts the expectation: the clients who have actually watched these tools work are the more sceptical group. Seventy-seven percent of those who have used consultants’ AI tools think the technology is a bubble, against fifty-five percent of those who have not. Familiarity lowered trust rather than raising it.

What it means for you. If you deliver work anyone pays for, the question is no longer whether to use AI. It is what you say about it, to someone better informed and less impressed than they were a year ago. Hiding it is not a strategy and confessing everything is not either. What the buyer is actually paying for is a named person who is answerable when the work is wrong.

What I would do this week. Write one sentence you would be comfortable saying out loud if a client asked directly. Not a policy. One sentence. If you cannot write it, that is the thing to fix before the question arrives.

Move 1: Write the four-answer memory page. What you sell and to whom, how you price, what makes a good client, and three settled decisions. One page, one hour, pasted into every session from now on. This is the whole fix for Leak One: the conversation has no authority in it, so you supply one.

Move 2: Move the constraint to the top. Whatever must always be true goes first, never buried mid-brief. This is free and it recovers most of that 30 percent.

Move 3: Break the marathon session. When it gets long, have it write down what has been decided, then start clean with that summary. Long sessions feel efficient and measure badly.

Move 4: Write your one sentence about AI use. Before a client asks, not during. Check it against the six things that should stay human.

None of these needs a developer and none needs a new subscription.

PROMPTS

The Four-Answer Memory Page. Walks away with a one-page description of the business that makes every future session specific instead of generic. Asks the four questions, produces the document, says where to keep it.

GO TO PROMPT

The Session Handoff. Walks away with a clean summary of everything decided in a long session, formatted so it can start the next one. Built for people whose conversations run all day and drift by afternoon.

GO TO PROMPT

The One Sentence. Walks away with a single line they can say to a client about how AI is used in their work, drafted from their own delivery process rather than a template.

GO TO PROMPT

Run in order they cover the same three leaks: what the business knows, what a session remembers, and what the buyer is told.

Operator Second Brain Starter Pack (Free Subscribers)

8 plain text files and a morning routine. That’s it. No software to install, no tool to buy, no architecture to design.

Drop these files somewhere your AI assistant can read them (ChatGPT projects, Claude projects, a local folder your AI tool has access to). Fill them in once. Update them as you go. Read them every morning with AI before you start working.

That’s the whole system. Operator Second Brain

These are notion text files. For our paid subscribers, you can access my Second Brain Prompts at the end of this edition.

Subscribe now

See how your whole business is organized. There are nine moving parts to a healthy business. One is under strain. Most operators fix symptoms instead of finding the real constraint. The Thrive Business System Assessment shows you all nine and helps you find which one matters first. Numbers & Decisions might be it. Or it might be something else. The map shows you. Start at learn.buildtothrive.co/build

Run the AI Leverage Assessment. If you already know the strain is operational drag, time leaks, or workflow bottlenecks, this is the faster path. It shows you the cost of friction and where AI actually helps. Start at learn.buildtothrive.co/mytimeback

The Thrive Business System Assessment. Run this diagnostic to see all nine moving parts of your business and surface which one is under strain. Most operators fix symptoms instead of finding the real constraint. This assessment walks you through each lever—Clarity, Opportunity, Follow-Through—and names where the bottleneck sits. Numbers & Decisions might be it. The Thrive Business System

For our paid subscribers

Read more

The strange part is that nothing broke. No tool or model failed me in a way I could point at, and I have more working parts running now than I planned to have by this point, and I am still behind the date I wrote down. For a few weeks I read that as a discipline problem, which is the read most of us reach for first, and it is usually wrong.

Here is what actually happened, and I think it happens to a lot of people building right now.

The week my own system refused to work

Three weeks running, one of my agents produced nothing.

It is the one that drafts the weekly newsletter. It fires, it reads the research, it checks the inventory, and then it stops and writes a short note that says it is waiting on me to choose the topic. I built it that way on purpose, back in July, after I caught myself reviewing finished editions I had never asked for and rejecting them. So I locked it, on the rule that it does not draft anything until I have chosen the topic myself.

The lock worked exactly as designed. Every week it read the research, found no decision from me, and halted.

What I had built was not a writing bottleneck. It was a mirror. The agent could do the work in minutes. It could not do the one thing the work depended on, which was me deciding what the week was about. Three weeks of a machine sitting still, waiting, while I told myself I was behind on production.

I was not behind on production. I was behind on decisions, and I had never noticed because until this year, my slow execution had been quietly absorbing them.

What speed actually exposed

For twenty five years across five businesses, the gap between deciding something and seeing the result of it was long enough to hide a lot of vagueness. I could say “we should go after that market” and spend six weeks building toward it, and somewhere in week four the decision would finish making itself. The work did the thinking. I did not have to be precise up front, because the time it took to execute gave me room to get precise on the way.

That room is gone. When execution collapses from six weeks to an afternoon, every decision I was half making now has to be whole before anything moves. The vagueness that used to dissolve inside the build has nowhere to go. It sits at the front and blocks the queue.

This is why the build feels slower even while everything measurable about it got faster. I am not producing less. I am hitting the unresolved decision sooner, and hitting it more often, and there is no longer a four week buffer where it quietly resolves itself.

You have probably seen a version of this number: roughly 95 percent of corporate AI pilots produce no measurable return. It comes from an MIT-linked report in August 2025, it has been argued with since, and I would not build a case on it. I mention it because when I read past the headline, most of those stalls are not the model underperforming. They are organizations that automated a process nobody had decided the shape of, so the tool arrived on schedule and the judgment about what it was supposed to do never got made.

I wrote something related last year about what happens to high performers when output stops being the scarce thing. I was looking at it from the employee side then. Running my own build, the same shift showed up from the owner’s chair, and it was less comfortable.

Three signs this is what is happening to you

Not every slow build is this. Sometimes a build is slow because the thing is genuinely hard. A few weeks ago I wrote about how to tell which broken thing in your business is worth fixing first. What follows is the answer I kept getting wrong in my own, for months, with that general version already written down. These are the signals that say the delay is sitting somewhere else.

One. Your output went up and your finished work did not. You are generating more drafts, more versions, more options than you were a year ago, and fewer of them are crossing the line. The pile in the middle is growing. That middle is where undecided things go to wait.

Two. Your tools are waiting on you more than you are waiting on them. Count the number of times this week you opened something, looked at it, and closed it without moving it. If the answer is more than a couple, the constraint is not capacity.

Three. You keep rebuilding instead of finishing. Going back to improve something already built feels like progress and asks nothing of you. Finishing requires a decision you have been avoiding. Rebuilding is the more comfortable of the two, and it is very easy to do for a month without noticing.

If two of those three are true, more speed will not help you. I tried that first. I added capacity to a system whose actual constraint was one person’s unmade decisions, which is a fast way to build a larger queue.

The thing I got wrong for about six weeks was assuming the delay was a capacity problem, so I kept adding capacity. What finally moved the build was going the other direction, and the specific change I made is small enough to describe in a paragraph.

The Decision Debt Audit

This is the exercise, and it takes under an hour. I run it Friday afternoons now.

Step one. List everything currently built and not running. Any process, agent, page, offer, sequence, or asset that exists and is not live. Do not judge it yet. Just get the list.

Step two. For each one, write the sentence that starts “This is waiting on.” Be specific and name a person. “This is waiting on me to decide whether we are still selling the thing” is a real answer. “This is waiting on more time” is not, and if that is what you wrote, you have not found the real blocker yet.

Step three. Sort into four piles.

• My call, and I can make it in five minutes. Make them now, in the session. Most of my eleven were here. This pile is where the build gets unstuck.

• My call, and I need information first. Write down the one piece of information and where it comes from. Not a research project, one specific thing. If you cannot name it in a sentence, the decision is probably in pile one and you are avoiding it.

• Not my call. Send it to whoever it belongs to today, before you close the session.

• No longer needed. Kill it. Some of what is waiting on you is waiting because it stopped mattering and nobody said so. Killing these is the fastest capacity you will find all week.

Step four. Count pile one. That number is your decision debt. Mine was seven the first time I ran it. It is usually one or two now, and the build moves at a pace I recognize.

What paid subscribers get in this piece

The rest of this is the part I could not have written six weeks ago, because I had not fixed it yet.

• The one change that unblocked the build, written out step by step so you can run it this week

• A worked example from inside Build to Thrive, including the three-week halt, what it cost me, and the two decisions that were actually behind it

• Where this goes wrong, the two failure patterns I hit while fixing it

• How to tell a real capacity problem from a decision problem, so you do not apply this to a build that genuinely just needs more hours

• The Loop Audit. Is the exercise I use to walk a single process end to end and find where it stalls. It pairs with this piece directly.

Paid subscribers also have access to my library of more than 200 tools and prompts.

It runs $80 a year, and you can add two hours of my time on an AI project of your choice by subscribing as a Founder.

The change that moved it

I stopped adding to the system and spent one session writing down every place it was waiting on me.

Not every task. Every decision. There is a difference, and the difference is the whole thing. A task is work I have not done. A decision is a judgment nobody else can make, which some piece of already built work is currently parked behind.

I found eleven. I had assumed there were two or three.

Read more

I have spent this year putting AI into my own business one area at a time, keeping track of which areas paid off and which wasted my time. The wins came from three ordinary parts of the business where I had been carrying everything in my head, and where writing it down once helped my scale my business in multiple levels. This is a topic I keep coming back as it has several benefits and it is foundational to any succesful AI operation. Two related pieces you might want to read talk about why the tools kept making me busier, and the value of loops and how I built the daily brief that runs my mornings where I tell that story of my Chief of Staff and how it made me so much more productive. In this particular issue I am going to walk your through three areas where AI pays off first.

Area One: your business remembers nothing. Positioning, pricing logic, how you handle a difficult client, why you dropped a service. Most of it lives in your head, and when it lives there, AI gives you generic output because it is working with nothing about you. New team members restart from zero. So do you, every time you sit down.

Area Two: follow-up runs on mood. Leads go quiet, proposals stall, past clients disappear. Not because anyone decided to drop them, but because following up depends on remembering, and remembering competes with everything else in your week. The fastest place to find money you already earned.

Area Three: the invisible work eats the week. Copying between tools, rewriting the same email, rebuilding a document you have built forty times. It never appears on a plan and it takes most of the day.

The three connect. What lives only in your head cannot be handed to anyone, human or otherwise, which is why the business still runs through you.

This edition will give you the right tips to make AI compound your memory. I hope you enjoy it.

Juan

68 to 77 percent. Small business AI adoption between April 2025 and January 2026, per Intuit’s 2026 AI Impact Report, with generative AI use climbing 40 to 58 percent. It went up, not down.

42 to 28 percent. The decline everyone is still quoting, from a survey fielded April 2025. Fourteen months old as a description of now.

$201.9 billion. Gartner’s 2026 figure for money going into AI that acts on its own, up 141 percent in a year. Spending runs well ahead of anyone checking what it produced.

1 to 15 people. The size of business this issue is written for. Most AI coverage assumes an IT department, which is why so little of it is usable.

95 percent. Big-company AI pilots that stall before going live, per MIT NANDA. Company, not evidence: August 2025, disputed, not peer reviewed.

If you like to know where your business is dragging and its costs, try my AI Leverage Assessment. It will help you diagnose how much it is costing you. If you like to chat afterwards about how to implement a solution, I help you get started with the right direction and tools. You can do it on your own (free), with me by subscribing to Build to Thrive’s founder subscription, or I can do it for you.

Subscribe now

Story 1: Your Business Remembers Nothing, So Your AI Sounds Like Everyone Else’s

What is happening. The most common complaint I hear about AI is that the output is generic. People assume they need a better tool or better wording. Usually the business has never written anything down. Your positioning, how you price, what makes a good client, why you stopped offering that one service. Real knowledge, all in one head.

What it means for you. AI can only work with what you give it. Handed nothing about your business, it returns the average of the internet, which is what generic means. This is also why the business cannot run without you. Knowledge that was never written down cannot be handed to a new hire, a contractor, or software.

What I would do this week. Open one document. Write four things: what you sell and to whom, what makes a client a good fit, how you decide pricing, and three lessons you would tell someone taking over tomorrow. One hour. Paste it at the top of your next AI conversation and watch the output stop sounding like a brochure. Highest-return hour in this issue.

Story 2: Follow-Up Is Where Your Money Is Already Sitting

What is happening. Leads go quiet. Proposals stall. Clients from two years ago, who liked you, have not heard from you since. Nobody decided to let those go. They went because follow-up depends on remembering, and remembering loses to whatever is on fire today.

What it means for you. This is money you already paid to earn. You ran the ads, had the conversations, did the work that made those people trust you. Letting it lapse means paying for the same lead twice. It is the cheapest revenue a small business has and the easiest to lose, because nothing breaks when you drop it. It just quietly does not happen. A warm relationship is also the shortest path to a paid short engagement.

What I would do this week. Put every open relationship in one list with three columns: who, what the next step is, and when. I run mine in three tracks by how warm they are, so the hot ones get worked first and nobody sits in an undifferentiated pile. Then have AI draft the follow-ups from your own notes and send them yourself. The drafting was never the hard part. The remembering was.

Story 3: The Number Telling You to Give Up Is Fourteen Months Old

What is happening. The stat circulating says small business AI use fell from 42 percent to 28 percent in a year. James Presbitero surfaced it in When To Not Build Your Own AI, which sent me to check where it came from. It traces to a survey fielded April 2025. Intuit’s report, running through January 2026, has adoption moving the other way: 68 to 77 percent.

What it means for you. If you tried AI, it did not stick, and you concluded the market was turning against it, the evidence you are using describes a window that closed over a year ago. Adoption did not reverse. The newer numbers suggest most people who quit were doing it alone, with nobody to tell them which part to start with. Different problem, different fix.

What I would do this week. Stop trying to fix everything at once. Pick the one area above where the pain is loudest and give it a single hour. Attempting all nine is how people conclude AI does not work for them. One area, finished, is how it starts paying.

THE PLAYBOOK

Move 1: Write the four-answer memory page. What you sell and to whom, what makes a good client, how you price, and three lessons you would hand your successor. One page, one hour. Paste it into every AI conversation from now on.

Move 2: Put every open relationship in one list. Name, next step, date. Sort by how warm they are. You will find revenue in the first ten minutes, because at least one person went quiet who was ready to buy.

Move 3: Name the task you have rebuilt forty times. The document, the email, the report you make from scratch every month. Write down how you do it, step by step, as if explaining to a new assistant. That written version is what AI can finally take over. Check it against the six things that should stay human before you hand any of it over.

Move 4: Do one area, not nine. Finish one before starting the next. The people who conclude AI does not work for their business are usually the ones who tried everything at once.

Not one of these needs a developer, and none of them need a new subscription.

HELPFUL PROMPTS

The Business Memory Page. Walks away with a one-page description of your business that makes every future AI conversation specific instead of generic. Answers four questions, produces the document, and tells them where to keep it. Open the prompt

The Quiet Deal Finder. Walks away with a list of every relationship that went quiet, ranked by how close it was to money, with a first follow-up line drafted for each. Built for people whose pipeline lives in their inbox and their memory. Open the prompt

The Repeat Work Audit. Walks away with the three tasks you rebuild most often written down as steps, so the work can be handed off. Names which of the three to hand over first and why. Open the prompt

Run in order they cover the same three areas: what your business knows, who it owes a reply, and what it keeps redoing by hand.

Share

HELPFUL TOOLS

• Scattered knowledge? Build memory. Start with the Memory Gap Audit.

• Get it →

• Inconsistent execution? Build workflows. Start with the Operations to Automation Map. 🔒Get it →

• Stalled decisions? Build decision support. Start with the Decide and Deliver prompt. Get it →

• Weak marketing? Build an authority flywheel. Start with the Authority Flywheel Audit. 🔒 Get it →

• Lost revenue? Build follow-up tracking. Start with the Warm Lead Re-engagement prompt. Get it →

• Inconsistent delivery? Build delivery support. Start with the Company Health Diagnostic. Get it →

• Founder dependency? Build a Chief of Staff layer. Start with the Morning Checkin 🔒. Get it →

To unluck the rest of the resources and access our full library of more than 200 prompts, skills and loops

Become a paid subscriber

Build to Thrive just hit #26 Rising in Business on Substack last week

Thank you for reading my newsletter and helping me grow at a pace of 600 subscribers per month. I started this newsletter in April of last year and it is now a resource to more than 5000 entrepreneurs. I would love for you to take the next step by offering a 25% discount over my paid subscription. By becoming a paid subscriber you have access to my library of prompts, workflows and timely content to help you gain your time back, improve your income and have happier customers and stakeholders. All for the price of a latte or capucchino. (Offer valid until August 15th

Subscribe now

A word on why I write it. I have started four companies, and I am building my fifth: this one, a media company I run almost entirely solo, on a fleet of about thirty AI agents that draft this newsletter, scan my inboxes, prep my client work, and write my morning brief. I am not reporting on this shift from the sidelines. I am living it, and I share what works and what breaks, openly, as I go.

If you want the right AI solution for your business, not just another list? Book a 15-minute call. I will assess your bottleneck and point you to the right system or app that fits. If it needs more digging, I will come back with the answer on a follow-up call.

Let's connect

TL;DR

Collecting AI tools won’t change your business. Building systems will. This piece shows you what an AI-powered business actually does that a pile of apps can’t: it remembers so you stop starting from zero, turns one-off prompts into loops that run without you, prepares your decisions so you spend judgment only where it counts, mines your content for what buyers actually want, and stops revenue from leaking in forgotten follow-up. The move is simple: find the one place your business is losing time, money, or momentum, and fix that first instead of buying another tool.

It comes with 8 tools built in, one for each way a business leaks leverage, so you can go from reading to fixing the same day

How to Build you AI-Powered Business

You have ChatGPT open in a tab right now. Maybe Claude too.

You use it to write emails, summarize the occasional document, or brainstorm a post when you get stuck. All of that is useful, but it may not change how the business operates. At the end of the week, the same work is still piling up, the same opportunities are slipping through, and ChatGPT has become a faster assistant sitting on top of an unchanged business.

If that sounds like you, you are not doing it wrong. You are just early. And you are asking the question almost everyone asks first: “What tools should I use?”

Natural question. Also a small one.

An AI-powered business uses technology to improve how the company remembers information, follows through on commitments, prepares decisions, and learns from its work. Writing an occasional email with ChatGPT may save a few minutes, but the larger benefit comes when AI becomes part of the way the business operates.

1. Stop chasing tools, start looking for leverage

A better place to begin is the part of the business that repeatedly wastes time, loses money, delays decisions, or allows opportunities to stall. Adding AI before understanding that problem may only accelerate an ineffective process. Identify what is breaking first, then decide what kind of system could improve it.

2. Build a memory your business can query

Most small operations have a memory problem. Important information is scattered across email, voice notes, calendar events, meeting notes, CRM fields, proposals, documents, DMs, and the operator’s head.

The consequences appear throughout the business. Important details from a prospect may be buried in an old email, a client request may never make it out of the meeting notes, and a useful audience question may disappear in the comments. Even strategic decisions can resurface months later when nobody records why they were made.

This is not an organization problem. It is a growth problem. A business without memory keeps starting from zero.

A useful business memory should make it easy to retrieve what a client cared about, what the company promised, which objections keep appearing, and which decisions have already been made. It should also show what is still unresolved and who is responsible for the next step.

This does not mean AI makes the decisions. It means the business remembers enough for you to make better ones.

3. Stop feeding the vending machine

Many people use AI like a vending machine: enter a prompt, collect the answer, and move on. That can save time, but the process disappears as soon as the task is finished.

There are three useful levels to understand.

A prompt handles a single request. You define the goal, evaluate the answer, and continue working. It is appropriate for one-time tasks such as drafting an email or summarizing a call, but it does not create a repeatable process.

A skill preserves the instructions behind a useful prompt so the same type of work can be completed consistently. Instead of explaining the task again each time, you solve the process once and improve it as you use it.

A loop, or workflow,, connects that skill to a trigger and a follow-through process. An event occurs, the skill prepares the work, the system tracks what should happen next, and you step in when the task requires judgment.

These levels build on one another. A useful prompt can become a skill, and that skill can later operate inside a loop. The prompt is not discarded. It becomes part of a more dependable system.

Consider a simple sales loop. A prospect completes an assessment, AI summarizes the responses and identifies the likely bottleneck, and a call brief is prepared for review. You lead the conversation, while the system captures the next steps and tracks the follow-up afterward.

The value comes from the full loop, not from any single prompt.

Our paid subscribers gain access to our full library of prompts, skills and loops you can

4. Protect your judgment, do not delegate it

Many operators already have enough information. The harder problem is deciding what to do with it.

A prospect may deserve another follow-up, but the previous conversations need to be reviewed first. A new product idea may sound promising, but it still has to be compared with existing priorities. Pricing changes, workshops, client engagements, and automation decisions all require context.

AI can prepare these decisions by summarizing the situation, comparing options, retrieving previous decisions, identifying missing information, and surfacing potential risks.

Some decisions should not be delegated. The final decision still belongs to the operator.

Pricing, positioning, client diagnosis, product promises, offer strategy, and major tradeoffs carry consequences that should not be delegated automatically. AI is most useful when it handles the preparation, allowing you to spend your judgment where it matters

5. Turn content into a learning loop

Content is usually treated as a marketing activity. Publish the post, send the email, look at the results, and start again.

An AI-powered business can use content as a source of market intelligence. Comments, replies, workshop questions, email responses, and sales objections reveal where people are confused, which problems keep appearing, what they want to learn, and how they describe those problems in their own words.

Most of these signals disappear because nobody captures them consistently.

AI can help collect and organize that information before it is lost. A useful comment may suggest an article, while several related questions may point toward a workshop. Recurring objections can improve an existing offer, and patterns from client work can support a case study, service, or digital product.

Once those signals are captured and reviewed, content becomes part of the learning loop. The business listens to the market, creates something in response, observes the reaction, and uses what it learns in the next cycle.

6. Plug the revenue leak in your open loops

A considerable amount of potential revenue disappears during follow-up. The problem is usually not a lack of interest or effort. It is the absence of a reliable system for managing open loops.

A prospect asks you to circle back in two weeks. A client requests something after a call. A partner suggests reconnecting the following month, or a reader asks a question that could lead to a useful conversation. Once the day becomes busy, those commitments are easy to forget.

An AI-powered business tracks its open loops. The system should show who is waiting for a response, what was promised, when the last contact occurred, and who owns the next action. It should also distinguish an active opportunity from one that has been parked, abandoned, or left overdue.

Many viable opportunities are lost because the next step depends on someone remembering it at the right moment. The prospect may still be interested and the relationship may still be developing, but the conversation stops because the open loop has no clear owner.

7. Make every engagement compound

Client delivery has consequences beyond the work completed during a single meeting. It affects trust, future decisions, and what the business learns from the engagement.

AI can support delivery by preparing for calls, summarizing sessions, recording action items, tracking progress, and identifying patterns across the work.

Before a coaching or advisory call, the system can assemble the client’s history, previous decisions, unresolved questions, and likely priorities. Afterward, it can record what was decided, what each person agreed to do, and which issues need to carry into the next conversation.

Patterns may also emerge across several clients. A problem that appears repeatedly could lead to a guide, workshop, productized service, or digital product.

The client benefits because fewer details and commitments are lost. The business benefits because each engagement contributes to future frameworks, offers, products, and content.

An AI-powered business compounds when delivery becomes part of a learning loop rather than an isolated transaction.

8. Add a Chief of Staff layer

As the amount of information grows, another report is unlikely to solve the problem. What the operator needs is a concise daily operating brief.

A useful AI Chief of Staff should identify what deserves attention, what may be at risk, and which commitments are waiting for a response. It should separate urgent work from tasks that can wait and highlight the one to three priorities that should not be displaced.

A dashboard presents data. A Chief of Staff brief interprets that information in the context of the workday and helps the operator decide where to focus.

The purpose is not to create more activity. It is to reduce the time spent sorting through information before meaningful work can begin.

9. Find your bottleneck

One part of your business probably felt more familiar than the others.

Perhaps important information is difficult to retrieve, or follow-up still depends on memory. Your content may be disconnected from your offers, while client delivery continues to depend too heavily on you. Decisions may stall because the necessary context is scattered, or the business may keep building from internal ideas without learning enough from customer signals.

Start with the problem that creates the greatest recurring loss of time, money, or momentum. Here is a tool I give clients to help them diagnose bottlenecks AI Leverage Assessment.

Instead of asking which AI tool to adopt next, identify the part of the business that needs to become easier to manage, more consistent, or less dependent on manual effort.

Once the bottleneck is clear, select the resource that addresses that specific problem. Some resources are available to all readers, while others are included with a paid subscription.

• Scattered knowledge? Build memory. Start with the Memory Gap Audit.

• Get it →

• Inconsistent execution? Build workflows. Start with the Operations to Automation Map. 🔒Get it →

• Stalled decisions? Build decision support. Start with the Decide and Deliver prompt. Get it →

• Weak marketing? Build an authority flywheel. Start with the Authority Flywheel Audit. 🔒 Get it →

• Lost revenue? Build follow-up tracking. Start with the Warm Lead Re-engagement prompt. Get it →

• Inconsistent delivery? Build delivery support. Start with the Company Health Diagnostic. Get it →

• Founder dependency? Build a Chief of Staff layer. Start with the Morning Checkin 🔒. Get it →

To unluck the rest of the resources and access our full library of more than 200 prompts, skills and loops

Become a paid subscriber

Want the right tool for your business, not just another list? Book a 15-minute call. I will assess your bottleneck and point you to the tool or system that fits. If it needs more digging, I will come back with the answer on a follow-up call

15-min tool delivery

This is the work behind Build to Thrive

This is why I do not start with AI tools when I work with a business. I start by finding the bottleneck.

AI is not the strategy. AI multiplies the way the business already works. If the process is unclear, AI produces unclear output faster. If follow-up is broken, AI sends more messages without building trust. If the offer is weak, AI writes better copy for a weak offer. If the business has no memory, AI keeps starting from zero.

But give it clear loops, clear decisions, clear ownership, and useful memory, and AI becomes leverage.

That is the work I am building at Build to Thrive: helping operators build AI-powered businesses that create more revenue, profit, and time.

Where to start

You do not fix this with another tool. You fix it at the bottleneck. The AI Leverage Assessment tells you which part of your business is losing the most time, money, or momentum, and which one to redesign first.

Take it here: Find your bottleneck

Let me know what you think of this content, I would love your feedback in the comments section.

Thank you reading. Feel free to share it!

Share

Oh…I almost forgot.

Want more free resources? Fifty operators shared the prompts behind the first six months of our AIBlueprint Edition. 50 Creators share their rockstar prompts →

Liked, downloaded and shared by more than 1000 people…you can’t go wrong!

In this edition you will walk away with one number that tells you whether your AI is an asset or a leak: cost per outcome. You learn why 70% of agents fail in production (no stop condition) and how to fix it. You get a simple follow-up rhythm that stops opportunities from disappearing. Three prompts you can use today: an outcome cost calculator, a stop-condition builder, and a follow-up rhythm designer. You understand why “my AI made more stuff” is not the same as “my AI made me money,” and the exact moves to close that gap before you spend another dollar automating.

Free mega resource for our readers

50 Creators Shared Their Rock-Star Prompts. Here They Are

EDITORIAL

Your AI made the output faster. That does not mean it made you richer. 93% of businesses report AI improved production. 57% still cannot show ROI that beats the spend. The gap is in one number: cost per outcome. This week we show you how to measure it.

Last week, we introduced the Thrive Business System: nine parts of a healthy business grouped into three clusters. This week we go deep on one: Numbers & Decisions. It is the lever that sits between “my AI made more stuff” and “my AI made me money.”

Here is what Numbers & Decisions means: the business can see what is actually working and make better decisions about what to fix next. When it is weak, the founder operates on gut feel. When it is strong, the numbers tell the story.

If you are an operator running AI work and seeing output up but margins flat: you need Numbers & Decisions working. If you are a consultant packaging AI-leveraged offers: you need it to price by outcome, not by input. If you are a founder automating a bottleneck: you need it to tell you whether the bottleneck was designed wrong first.

Three practical questions close the gap: What did the AI work cost? When should the loop stop? Which opportunities are worth following up on?

The numbers that tell you whether your AI is an asset or a leak.

Juan

BY THE NUMBERS

The gap between AI output and AI return is measurable. Here is what that gap looks like in five numbers.

1. 93% of businesses report AI improved their output. 57% still cannot show a return that beats the spend. The gap between more output and more return is the whole game. Domino Data Lab, State of AI 2026 — verify exact figures before publish

2. Agent workflows cost 19 to 50 times more than a single model call. The variance comes from context window size, retrieval complexity, and how many steps the agent needs to take. One run might cost $0.10. The next run on the same task costs $5. Pricing off an average misses the gap entirely. Atlan

3. In cost breakdowns for AI workflows, model tokens typically represent 8 to 27% of total expense. Infrastructure, retrieval, storage, and human review make up the rest. Optimizing tokens alone leaves 73% of cost unaddressed.

4. Agent systems in production often encounter reliability issues when designed without clear stop conditions and validation gates. The problem is usually not the model. It is that the operating loop was never designed with reliability in mind.

5. Most sales opportunities are not lost to rejection. They are lost to no follow-up. A simple rhythm beats a complicated CRM nobody uses.

SIGNAL

Output and Return Are Not the Same Thing

The gap is real, and it is expensive. While most organizations report AI improved their output, fewer can demonstrate that AI improved their return. There is a measurement gap, and it is costing money.

Why it matters: Faster work does not automatically mean richer work. Your AI made the task faster. That does not mean it made you more money. The gap sits in Numbers & Decisions, the lever that shows what is actually working. Most operators never measure it.

What I am observing: The operators closing this gap first do one thing different. They price AI-leveraged work by outcome, not by input. Petri Salonen has written about this: the cost per run varies wildly depending on model load, retrieval complexity, and what the AI actually uses. Pricing off the average is how you automate yourself into a loss. The fix is not faster AI. It is outcome-based pricing.

Stop Conditions Come Before You Scale

Many AI agents fail in production. Not because the model is weak, but because the operating loop was never designed with reliability in mind.

Why it matters: An agent without a stop condition is a loop without an exit. It will run until it costs you. The real fix is not better prompts. It is better design. Two things matter: an eval gate that asks “is this output worth keeping” and a stop condition that asks “should we keep going.” Those happen before you let AI run it at scale.

What I am observing: The operators building reliable systems do this backwards from the usual AI workflow. They design the operating loop first. They test the stop conditions with small batches. Then they let AI run it. The difference between that approach and “just turn on the agent” is the difference between “output worth keeping” and “output that costs you.”

Story 3: Relationships Die in Silence, Not Rejection

Most opportunities disappear because nobody followed up. Not because the opportunity was bad.

Why it matters: An opportunity without a rhythm becomes an opportunity that gets lost. A follow-up rhythm is simpler than it sounds: who owns it, when does it happen, what is the next step. The business keeps it moving.

What I am observing: The operators who win this stop buying CRM software and start designing a rhythm. AI can help send the message, summarize the call, surface what is stale. But AI only helps after the rhythm is designed. Bolt AI onto scattered follow-up and you get faster scattered follow-up.

THE PLAYBOOK

Measure outcome cost, not input cost.

You need one number that tells you whether the AI work paid for itself. Build that number this week. It rewrites how you price and what you automate next. The number is cost per successful outcome, not cost per token or per run.

Design the stop condition before you scale the loop.

Ask: when should this agent stop running? When the output is unreliable. When we have spent enough. When the job is done. Design that exit before you let it run. Then the loop is worth multiplying. More on loops in this article: Why Your AI Is Making You Busier: The 6-Part Framework for Real Delegation

Move 3: Own the follow-up rhythm.

Pick one day a week. Pick who owns it. Pick what “next step” means. The rhythm keeps the relationship moving. Everything else is noise.

Move 4: Let the numbers decide what to fix next.

Your business has multiple moving parts working together. Most operators fix the one that screams loudest. The numbers show you which one actually holds everything back. Fix that first. That is the difference between activity and leverage.

Three prompts to measure and design what works this week.

You walk away with a spreadsheet that shows the real cost of one AI-leveraged workflow. Input: task description, model used, success rate, cost per run. Output: cost per successful outcome and the threshold for profitability. This is the number that changes your pricing.

GO TO PROMPT

You walk away with a two-part test for any agent: (1) how do we know this output is worth keeping, (2) when do we stop the loop. Without these, the loop runs until it costs you. This is how reliability gets designed in.

GO TO PROMPT

You walk away with a repeatable rhythm: cadence (weekly, bi-weekly), owner (who is responsible), next step (what moves the relationship forward), and how to surface what is stale. This is how opportunities stop dying in silence.

GO TO PROMPT

The three prompts are the three redesigns that come before you multiply with AI. Cost clarity, loop reliability, and relationship rhythm. Together, they close the gap between output and return.

B2T Prompt Vault — Own the prompts you use instead of chasing new tools. This week’s three prompts live here: outcome cost calculator, stop condition builder, follow-up rhythm designer. These prompts work inside Claude, ChatGPT, Notion, or your internal company drive. Save them. Use them. Build your library. Learn.buildtothrive.co/vault

Petri Salonen — Business Models in the AI Era. Salonen names the exact problem: the cost per run varies wildly, and pricing off averages fails. His work is the proof that Numbers & Decisions is the lever holding back ROI for operators scaling AI work. Read his latest on Business Models in the AI Era

See how your whole business is organized. There are nine moving parts to a healthy business. One is under strain. Most operators fix symptoms instead of finding the real constraint. The Thrive Business System Assessment shows you all nine and helps you find which one matters first. Numbers & Decisions might be it. Or it might be something else. The map shows you. Start at learn.buildtothrive.co/build

Run the AI Leverage Assessment. If you already know the strain is operational drag, time leaks, or workflow bottlenecks, this is the faster path. It shows you the cost of friction and where AI actually helps. Start at learn.buildtothrive.co/mytimeback

The Thrive Business System Assessment. Run this diagnostic to see all nine moving parts of your business and surface which one is under strain. Most operators fix symptoms instead of finding the real constraint. This assessment walks you through each lever—Clarity, Opportunity, Follow-Through—and names where the bottleneck sits. Numbers & Decisions might be it. The Thrive Business System

Thank you for reading.

Juan

Some news I want to share before you go.

Build to Thrive just hit #26 Rising in Business on Substack last week

Thank you for reading my newsletter and helping me grow at a pace of 600 subscribers per month. I started this newsletter in April of last year and it is now a resource to more than 5000 entrepreneurs. I would love for you to take the next step by offering a 25% discount over my paid subscription. By becoming a paid subscriber you have access to my library of prompts, workflows and timely content to help you gain your time back, improve your income and have happier customers and stakeholders. All for the price of a latte or capucchino. (Offer valid until July 31st)

Subscribe now

Thank you for reading my newsletter and helping me grow at a pace of 600 subscribers per month. I started this newsletter in April of last year and it is now a resource to more than 5000 entrepreneurs. I would love for you to take the next step by offering a 25% discount over my paid subscription. By becoming a paid subscriber you have access to my library of prompts, workflows and timely content to help you gain your time back, improve your income and have happier customers and stakeholders. All for the price of a latte or capucchino. (Offer valid until July 31st)

Subscribe now

TL;DR

Your business may feel broken because you are fixing symptoms instead of the real constraint. The Thrive Business System helps you find the weak point across 9 business levers grouped into Clarity, Opportunity, and Follow-Through. Traditional organizations add people, tools, and processes to manage complexity. AI-native businesses first redesign how workflows, then use AI to multiply what works. The goal is a business that earns more, runs smoother, and depends less on you.

Most founders do not wake up thinking, “My business model has a structural constraint.”

They wake up thinking:

Why am I still involved in everything?
Why are good opportunities not moving?
Why are clients happy with the work, but the delivery still feels messy?
Why did we buy another tool and still end up with the same problem?

That is usually how business problems show up.

Not as a clean diagnosis.
As friction.

A lead goes quiet. A proposal stalls. A client asks for an update that should have already been sent. A team member waits for the founder to make a decision. A new AI tool gets added, but the work does not really get better.

So the instinct is to fix what is loudest.

Work harder.
Hire someone.
Buy software.
Add automation.
Try another AI tool.

For a while, that feels like progress.

Then six months later, the same strain comes back in a different form.

That happens because the problem you noticed first was probably not the real problem. It was a symptom.

And this is the lesson I now lead with:

You cannot create leverage by speeding up a broken process.

Bolting speed onto chaos is not leverage

AI has made this problem more urgent.

A lot of businesses are not redesigning the work. They are just automating the work they already have.

That sounds efficient until you realize the process being automated should not exist in its current form.

If follow-up is scattered, AI can help you send messages faster. But the deeper issue may be that nobody owns the relationship pipeline.

If content feels generic, AI can help you publish more. But the deeper issue may be that the business has no clear point of view, no buyer language, and no visible proof.

If client delivery feels chaotic, AI can summarize meetings and draft updates. But the deeper issue may be that the client journey was never designed.

Speed helps only after the work is worth multiplying.

Otherwise, you do not get leverage. You get a faster mess.

The three questions I use to read a business

When I look at a business, I do not start with tools.

I start with three questions:

Clarity: Do we know what matters?
Opportunity: Can we create the right opportunities?
Follow-Through: Can we follow through consistently?

Every company is answering those questions all the time, whether the owner has named them or not.

When a business feels heavier than it should, one of those questions is usually weak.

The Thrive Business System breaks those three questions into nine parts.

Not as a checklist.
As a diagnostic map.

The point is not to label the business. The point is to find where the real strain is coming from before you spend more time, money, or energy fixing the wrong thing.

Clarity: Do we know what matters?

Clarity is where the business remembers, focuses, and decides.

Without clarity, everything downstream gets harder.

1. Business Memory

Business Memory is the shared context your company runs on: positioning, offers, clients, decisions, voice, lessons, processes, and operating knowledge.

When this is weak, everything lives in the founder’s head.

The founder repeats the same explanations. Client context gets lost. Content sounds generic. Every new project starts from scratch.

This is also why AI often disappoints. If the business has no reliable memory layer, AI has nothing specific to draw from. It produces generic output because the business gave it generic inputs.

2. Priorities and Rhythm

Priorities and Rhythm is how the business decides what matters now and keeps it moving.

When this is weak, everything feels urgent. The founder becomes the traffic controller. Important work gets buried under reactive work. Projects start, stall, restart, and then get replaced by the next urgent thing.

This is where founder dependency often lives.

Not because the founder is bad at delegation, but because the business has no rhythm that allows the right work to move without constant intervention.

3. Numbers and Decisions

Numbers and Decisions is how the business sees what is working and what is slipping.

This does not need to be complicated.

For a solo operator, it might be a simple view of leads, sales, delivery load, time leaks, and cash. For a larger business, it may include margins, financing, capacity, client experience, and capital allocation.

When this is weak, the business runs on instinct.

Revenue may be growing while margin is shrinking. Clients may be satisfied while delivery is becoming too heavy. The founder may feel busy without knowing which activity is actually creating value.

Clarity creates direction.

Without it, you may be working hard in the wrong direction.

Opportunity: Can we create the right opportunities?

Opportunity is where the business understands the market, earns trust, and makes it easy for the right people to take the next step.

4. Market and Client Insight

Market and Client Insight is how the business listens.

When this is weak, messaging runs on assumptions. Offers are built from the inside out. Content does not connect. Sales calls reveal the same confusion over and over.

The business misses what buyers are actually saying because nobody is capturing the questions, objections, language, and signals already coming from the market.

A business grows stronger when it listens carefully.

5. Reputation and Visibility

Reputation and Visibility is how the market comes to see what you know and why you are credible.

A lot of operators have real expertise that the market never sees.

They are good at the work, but invisible to the people who would value it. Referrals depend too much on personal relationships. Content is inconsistent. Buyers do not understand why they should trust them before the first conversation.

Visibility is not just posting more.

It is turning experience, proof, ideas, and perspective into public assets that make trust easier to build.

6. Offers and Next Steps

Offers and Next Steps is how the business turns interest into action.

When this is weak, people may respect you but still not know how to work with you.

Calls end without a next step. Proposals are too custom. Pricing feels improvised. Content creates attention but does not create movement.

A good offer makes the next step feel obvious and valuable.

Opportunity creates demand.

But the order matters. Visibility without clarity just broadcasts a confused message more loudly.

Follow-Through: Can we follow through consistently?

Follow-Through is where the business keeps relationships moving, delivers well, and reduces the friction behind the scenes.

7. Relationships and Follow-Up

Relationships and Follow-Up is how the business keeps important conversations alive.

When this is weak, leads go quiet. Proposals stall. Past clients are forgotten. Referrals never get followed up. Good opportunities sit in inboxes, notes, or memory with no next step.

Most opportunities are not lost because of rejection.

They are lost because no one followed up.

A simple follow-up rhythm is often more valuable than a complicated CRM nobody uses.

8. Client Experience

Client Experience is how the business delivers in a way that feels clear, professional, and trustworthy.

When this is weak, the final work may still be good, but the process feels improvised.

Onboarding is unclear. Updates are inconsistent. The client has to ask what happens next. The founder keeps stepping in to rescue the experience.

Great delivery makes clients feel guided, not managed.

9. Back Office and Admin

Back Office and Admin is the invisible work that keeps the business running.

Documents. Handoffs. Inboxes. Calendars. Internal requests. Repeatable tasks. Small administrative steps that do not look strategic until they start consuming the capacity needed for growth.

When this is weak, tools pile up. Processes go undocumented. Work gets copied manually. The business feels heavier than it should.

Many “people problems” are actually design problems.

The issue is not always the team member. Sometimes the work was never structured for delegation.

Follow-Through creates trust, delivery, and results you can repeat.

The loop matters

These nine parts are not separate boxes.

They move in a loop.

Clarity creates direction.
Opportunity creates demand.
Follow-Through creates trust, delivery, and repeatable results.

Then the loop feeds itself.

Follow-Through produces learning.
Learning improves Clarity.
Better Clarity creates better Opportunity.
Better Opportunity requires stronger Follow-Through.

That is why fixing one visible problem often fails.

A business weak in Clarity cannot always be repaired with more marketing. A business drowning in delivery issues may be carrying a weak offer, unclear priorities, or missing business memory. A business struggling with sales may not need a louder funnel. It may need better follow-up, clearer proof, or a simpler next step.

The map helps you see not only where the strain is, but where it may be coming from.

Where AI actually fits

AI is not a tenth part.

It is not the strategy. It is not the operating system.

AI is what you apply after you understand the business, find the strain, and redesign the work underneath it.

The sequence matters:

Understand the business.
Find the part under the most strain.
Redesign the work so it is worth doing.
Then use AI to improve the outcome, speed, quality, or consistency.

Run that sequence backward and you get automation without leverage.

Leverage is the mechanism. A better business and a better life for the person running it are the outcomes.

A real example

I learned this the hard way.

I inherited a management-intensive Section 8 housing portfolio. It had high friction, high founder dependency, thin margins, and too many moving parts.

The instinct would have been to work it harder.

More oversight.
More chasing.
More people solving the same recurring issues.

But the real strain was not demand.

The strain was in the operating model itself: Priorities, Follow-Through, and the way the work was structured.

So we redesigned the work and repositioned the portfolio into scalable middle-market multifamily.

Over time, that led to 134 units and a 112-key hotel repositioned, more than $8 million in investor capital raised and leveraged 4x, and 20 percent plus IRR across the portfolio. The work was later recognized by the Greater New Haven Chamber of Commerce. More on this story

I did not have this map written down at the time.

I have spent the years since turning what worked into something repeatable, so the next operator does not have to learn it one expensive mistake at a time.

How to use this map this week

You do not need a consultant to start.

Ask yourself three questions:

Do we know what matters?
Can we create the right opportunities?
Can we follow through consistently?

One will feel shakier than the others.

That is your cluster.

Then ask: which part inside that cluster is carrying the most weight?

Not the loudest problem.
Not the easiest problem to fix.
The one everything else seems to be waiting on.

That is usually the first-place worth redesigning.

Where this goes next

The Thrive Business System Assessment is the main doorway into this map.

It is built for the founder or operator who can feel that something is off but does not yet know where the real constraint sits.

It asks nine plain questions, one for each part of the system, and shows what is strong, what needs work, what is at risk, and which area deserves attention first.

If you’re open to it, I’d be happy to offer you a complimentary 15-minute virtual coffee. We can talk through your results, and I’ll recommend an AI tool you can use right away to address your specific challenge.

My only ask is this: if you find that after you apply what I said in our conversation you gain back 5 hours of your time, consider recommending me to someone who could also benefit or schedule a follow-up conversation with me to explore how I may be able to help you further.

If you already know the issue is operational drag, time leaks, manual processes, workflow bottlenecks, customer friction, or AI leverage, the AI Leverage Assessment may be the better starting point.

If you already know the issue is authority, positioning, content, offer clarity, public assets, or turning expertise into trust, the Authority or Value Clarity path may be the better starting point.

And if you already know you want to talk, book the free 15-minute call.

The Thrive Business System Assessment is the central map, not the only entrance.

However you enter, the sequence is the same.

We look at the business.
We find the part under the most strain.
We identify one useful move.
Then we decide whether there is a larger system worth building.

You cannot create leverage by speeding up a broken process.

You redesign the work first. Then you let AI multiply what is worth multiplying.

That is how you build an AI-powered business part by part, without betting everything up front.

If one cluster stood out as you read this, start there.

Or take the Thrive Business System Assessment and see your whole map.

Either way, the goal is the same: find the part that is actually holding the business back before you spend more time, money, or energy fixing the wrong thing.

Thank you for reading,

Juan

Wait…I want to leave you with something extremely valuable. A post that has more than 15,000 views and over 1300 downloads.

50 Free Resources that can take your Business to the Next Level

50 brilliant creators contributed to this.

Broken down into 9 topics:

• Leadership and Decision-Making: making the call you already have the information to make

• Content and Storytelling: shaping a piece of writing that proves the point

• Product and Building: shipping the minimum version of the thing

• Growth and Marketing: finding and moving the next customer

• Operations and Systems: turning a one-off into a repeatable process

• Research and Synthesis: making sense of what you read

• Career and Personal Brand: naming your value in a sentence a buyer can act on

• Creativity and Craft: stress-testing your own thinking

• Mindset and Operating: getting your head right to do the work at all

I really hope you can experience real progress and clarity out of these resources.

Build Leverage. Thrive on your terms.

Juan

Subscribe now

I just got back from Iceland, and I still cannot properly describe what I saw.

Someone asked me what surprised me most, and I meant to give one answer. I gave seven.

• The light. It never went dark, not once in a week.

• The weather has a temper: sun, sleet, and dead calm inside the same hour.

• I counted about a dozen waterfalls, then gave up counting.

• The mountains have no peaks. Most are flat on top, sliced clean, table after table.

• How far you can see. On the flat stretches it is a full 360, nothing but distance, until a wall of mountain simply begins.

• The earth is black. Black soil, black sand, black beaches.

• And then, somehow, green everywhere, against all that black.

That is Iceland. It over-delivers on every promise the photos make.

But the country was never really the story. The story is the six of us: me and my wife, our two eldest (both past eighteen now, which still catches me off guard), and the younger two, fourteen and eleven. We saw puffins, swam on hot springs, learned about lava and the power of volcanoes, geysirs, black beaches, more than 20 waterfalls, hiked, went of a mega zip line (1 Km) and many more memorable events. And here we are doing the same little dance in nine different places.

We love traveling while been in contact with nature and I know exactly what that age spread means. An eleven-year-old and a twenty-year-old are not on the same trip. One wants the hike and wants it now. One would sleep until noon given the chance and no moral instruction. One has been looking forward to the show for a month. And one asks how long the drive is roughly every eleven minutes, with complete sincerity, as though the answer might have improved. I wanted this trip to be good for all six of them, not good on average with two people quietly enduring it.

It started with one prompt

Here is how it began, and I bet it is how you begin too. Months out, I opened an AI and typed something like this: give me the top 20 things to do in Iceland on a seven-day trip for a family of six, ages 22, 19, 14, and 11. We like light hikes. We do not want to drive more than an hour on any single leg. We like to eat in affordable places, so add a couple of cheap options too.

Then I did the thing that changed everything. I let it look at my email. It read the booking confirmations it had already helped me put together, the flights, the rental car, the lodging, the airport parking, and it planned the whole trip around them. It was not guessing where we were staying or when we landed. It knew.

It knew other things because I told it. It knew we wanted to watch the World Cup, so it shaped each day so we would be back at the lodging in time for kickoff, and when the final came around it suggested a few bars in Reykjavík where we could watch it. It knew Iceland’s weather turns on you without warning, so it built a Plan A and a Plan B for the days that might get rained out.

And it did not stop at a list. For every stop it gave me the drive time and how long to actually spend there, a short description of what to do when we arrived, where to get the best ice cream, which local traditions and events were worth catching, and, because we like the markets where locals actually shop, it pointed us to a flea market.

Then I made it something we could see

A list in a chat window is still a list. So I turned it into a brochure, a proper document with a picture of every site, so the family could see where we were going and not just read the names. That was the first time the trip felt real to them.

But a brochure is fixed the moment you make it, and on the ground you need to know what today looks like. A document cannot tell you that. So I kept going and turned it into an app.

The app, tab by tab

The whole point was simple. Any one of the six of us could open the web on a phone and know exactly how the day was planned and what to expect. That mattered more than I thought it would, especially for the older two, who wanted to see what was coming next. It built excitement. It gave everyone the same picture of the trip. For once, we were all on the same page.

Here is how it looks.

Days. The whole week, one day at a time. Each day opens with where we are sleeping that night, then a timeline of stops, each with a time, what to do there, and a tap straight through to Google Maps. The most-asked question of any family trip, what are we doing today, finally had a five-second answer that did not go through me.

Stays. Every place we slept, with the address, the host’s number, the check-in time, and whether it was already paid. These were the exact bookings the AI had read out of my inbox, now sitting where anyone in the family could find them.

Do. The activities, and next to each one, what it actually costs per person, adult and child. This is where the light-hikes-only and affordable rules show up, with an honest per-person total at the top so there were no surprises.

Eat. Where to eat without turning every meal into a conversation about money, which in Iceland takes some doing. Affordable spots, the famous hot-dog stand, the farm that does the ice cream, each with a price estimate so we could choose before we sat down.

More. Everything you only need once but are grateful to have on hand: emergency numbers and the official road and volcano pages, a before-you-go checklist that remembers what you have ticked off, the World Cup schedule with the bar booked for the final, and the airport parking details for the day we flew home.

Ask AI. And then the part that pulled everything together.

The second brain

Here is the idea in one line. I stopped asking AI about Iceland and started giving AI my Iceland. Because it held the whole plan, it could answer the specific questions, like what time we had to leave to make something, and the general ones, like where to find puffins after we struck out, going to the web when the question needed the web. Same assistant everyone has. This one had read our trip.

At Þingvellir we had a whole national park in front of us and Geysir still to reach before the day ran out. I asked it which two things to do if we only had time for two. It gave me the gorge walk between the continents and the waterfall at the end of it, and pointed out they sit on the same path, so we got both in one walk and still made Geysir on time.

The day the plan broke, we had pinned our puffins on a clifftop stop and saw none. The useless question is where do you see puffins in Iceland. The one I asked was, we missed them, what is our real shot today, given where we are and what is still booked. One answer. A boat, a time, a place.

The difference was never that this AI was smarter. It was that it finally knew the one thing every other AI did not: our trip. That is the unfair advantage. Not an AI that knows more about Iceland than you do, one that knows your Iceland. Everyone else is asking a stranger for directions. You are travelling with someone who has already read the whole itinerary.

It is still there

Now that we are home, the live parts are quiet, but the whole trip is still on our phones, exactly as we did it, every day, every place, the map. It is not really an app anymore. It is the photo album that happens to be shaped like an itinerary.

I am shipping this early, and I want your feedback

One honest thing before I point you at it. This is early, and I am shipping it on purpose. The offline version and the deeper booking smarts are not done yet. I would rather put it in real hands, on real trips, and find out what breaks than polish it in private for another six months.

So here is the real app from our trip: iceland-itinerary.replit.app. Open it, walk the week day by day, and if Iceland is on your list, treat it as a starting map. Take the stops, the drive times, the picnic trick, and make your own notes. Then tell me what you would want for your own trip: what is missing, what you would change, what would make you actually use one of these.

I am improving it as I go, and I am honestly testing whether there is a business here at all. Your feedback is how I find out.

If you like me to help you with your next trip and build an app for you, or you want to know more about Iceland, feel free to reach out to me. I would love to talk to share with you my experience.

Subscribe now

Visit juansalasromer.com if you like to learn more about my work and how I can help

As a token of appreciation to my readers, I am discounting my paid subscription by 25%, only $5 dollars a month takes you to timely information and tools to gain time, money and happy customers. (Offer valid until July 31st)

Subscribe now

A few weeks ago I published a piece with 50 prompts from 50 creators. It was a compilation of prompts I accumulated over the first six months writing my Monday edition, The AI Blueprint, built with 50 content creators on Substack.

It became the most viewed article I have ever posted. More than 15,000 views in three weeks, and it keeps climbing. Here is the article if you have not seen it. 50 Creators Shared Their Rock-Star Prompts. Here They Are

Curious which prompt people clicked most, I was surprised the premortem prompt came out on top. It asked you to imagine a big bet had already failed, then work backward to figure out why.

That caught my attention because it says something about how people are actually using AI. They want to move faster, but they also want to feel less blind.

The most useful AI prompt might be the one that slows you down

Most AI content still orbits speed. Write faster, summarize faster, research faster. Speed matters, but speed is not judgment, and you can move quickly in the wrong direction. You can automate a broken process, launch an offer nobody wants, or build a funnel that earns attention and builds no trust. You can install AI into a workflow that nobody actually changes.

I see it with the clients I work with. They are moving confidently toward the wrong thing.

That is why the premortem is useful. It does not ask how to execute faster. It asks a more uncomfortable question: assume this failed, what happened? That question changes the room. It gives people permission to name the thing they were already worried about.

What a premortem is

A premortem is the opposite of a postmortem. A postmortem happens after something fails, and the team looks back to ask what went wrong. A premortem happens before the failure. You imagine the project, offer, or decision already failed, then you ask why.

The value is clarity, not pessimism. Most planning conversations are biased toward confidence. The founder wants momentum, the team wants direction, the consultant wants the client to say yes, the creator wants the launch to land. So the conversation drifts toward “here is why this can work.” A premortem creates a different conversation: here is how this could break. That does not kill the idea. It improves it.

Smart people fail for predictable reasons

A lot of capable people fail for reasons you could have named in advance. They commit too early to an assumption they never tested.

The offer sounds good, but the buyer does not care enough. The content gets attention, but creates no demand. The AI system is powerful, but the team never uses it. The product works, but the onboarding confuses people. The partnership looks strategic, but the incentives are misaligned. The service sells, but delivery becomes too custom to carry. The business grows, but the operator becomes the bottleneck.

These are not random. They are failure paths. Once you can name the failure path, you can design a better test.

In case you are interested in more prompts, the prompt navigator (The Vault) provides you with more than 200 prompts and tools to leverage your capacity and enhance your cognitive abilities. It is on a paid subscription service and it grows every week. Now running a 25% discount ($5 dollars a month)

Prompt Navigator

The real value is not the list of risks

A weak premortem hands you a long list of things that could go wrong. Market risk, execution risk, budget risk, competition risk. Mildly useful, not actionable. A strong premortem gives you the one or two assumptions that matter most.

If you ask AI “what could go wrong with this idea?” you get a generic list. That is the wrong question. The better ones sound like this:

• What would have to be true for this to work?

• Which of those assumptions is most fragile?

• What would we see early if it started to fail?

• What is the smallest test we can run before we commit more time, money, or reputation?

That is where the value lives. A premortem is a decision tool, not a fear exercise.

Premortem is a family of tools

The more I worked with this, the more I saw that “premortem” is not one move. It is a family of them, and the specialized versions are the ones I run most. Each is more than a question. It names the moment to run it, the failure scenario to sit with, and the diagnosis to work through before you commit. I provide links to these prompts at the end of this article.

• The Offer Premortem. The one that has saved me the most money, because many offers fail before the sales call. It is also the one I have built out the furthest, so it is worth showing in full (see link at the end of the article).

• The Founder’s Premortem. “Six months from now, this bet failed. Why?” Run it before any major launch, pivot, hire, or partnership to find the single point of failure before it gets expensive.

• The Funnel Premortem. “People clicked, read, and disappeared. Where did trust break?” Run it before you send traffic to a new lead magnet or sequence, because traffic is not demand.

• The Content Premortem. “The post performed, but created no business value. Why?” Run it before you commit to a series, to tell applause apart from progress.

• The AI Adoption Premortem. “We built the AI system, but nobody used it. Why?” Run it before you roll a tool out to yourself or a team, and design for behavior, not for capability.

• The Client Delivery Premortem. “The client paid, but the engagement went sideways. What happened?” Run it before kickoff, while the scope is still soft.

• The Scaling Premortem. “Revenue grew, but the business got heavier. Where did the model crack?” Run it before you add clients, channels, or headcount.

• The Partnership Premortem. “The partnership looked strategic, but it stalled. Where did the incentives split?” Run it before you sign, while the terms are still open.

Why AI makes premortems more useful

Premortems are not new, but AI makes them easier to run, and that matters. Most people do not have a room of advisors on call every time they make a decision. They may not have a partner willing to challenge every assumption, and even when they do, social pressure gets in the way.

AI can help here, not as an oracle or the final decision-maker, but as a structured thinking partner. It can simulate objections, classify risks, ask follow-up questions, surface hidden assumptions, and force specificity. It turns vague anxiety into testable concerns.

The point is not to ask AI “will this work?” The point is to ask “where am I most likely fooling myself?”

A premortem prompt you can use right now

Here is a basic version to try today:

Assume this initiative failed six months from now.

Here is the initiative: [DESCRIBE THE BET, OFFER, PROJECT, AI BUILD, FUNNEL, HIRE, OR STRATEGY].

Analyze the five most likely reasons it failed. For each one, identify:

1. The hidden assumption that turned out to be false

2. The early warning sign I should watch for

3. The cost if I ignore it

4. The smallest test I can run this week to reduce the risk

5. The one change I should make now before committing more time or money

Then identify the single biggest point of failure and explain why it matters most.

It works because it does not stop at “what could go wrong.” It pushes toward action. A good premortem leaves you with a better next move, not a longer worry list.

Do not treat every risk the same

One reason risk planning fails is that everything gets flattened into another bullet point. But some risks are unlikely and catastrophic, some are likely and manageable, some are reversible, some are fatal, some are emotional distractions, and some are true constraints. A good premortem sorts them by four questions: how likely is this, how costly would it be, how early could we detect it, and how easily could we reverse it. The goal is not to eliminate risk. It is to avoid being surprised by the obvious one.

The best operators are more specific

I think this is why the premortem prompt drew so much interest. It gave people a way to be cautious without feeling stuck.

There is a real difference between vague fear and specific risk. Vague fear says “I do not know, something feels off.” Specific risk says “if buyers do not already believe this problem is urgent, this offer will need too much education to sell at this price.” One creates hesitation. The other gives you a test to run.

That might be one of the most valuable ways to use AI right now. Not to replace your judgment, but to sharpen it before the cost of being wrong goes up.

Before the next big bet

Before you launch the offer, rehearse the failure. Before you build the AI system, rehearse the adoption problem. Before you publish the content series, rehearse the conversion gap. Before you scale the service, rehearse the operational drag. Before you say yes to the partnership, rehearse the incentive conflict.

Not to talk yourself out of ambition. Ambition gets stronger when it has been pressure-tested. The best ideas do not need less scrutiny, they need better scrutiny, earlier. A premortem lets you see the crack before you build on top of it, and use AI as a judgment tool, not just a productivity one.

In a world where everyone is trying to move faster, that might be the more valuable advantage.

Get the Founder’s Premortem Kit and my 7 go to premortem prompts. The kit includes the main premortem prompt, five specialized versions (offer, funnel, content, AI adoption, and client delivery), and a one-page decision brief you can use before any major bet.

Premortem Prompts

Founder’s Premortem Kit

As a token of appreciation to my readers, I am discounting my paid subscription by 25%, only $5 dollars a month takes you to timely information and tools to gain time, money and happy customers. (Offer valid until July 31st)

TL;DR — Authority Edition

Now that almost everyone has the same AI tools, knowing your field is no longer what sets you apart. What does is a system that turns what you already know into something people can see, trust, and act on. Think of authority as a flywheel rather than a stack of posts: the research, talks, and results you already have feed one clear idea, that idea becomes real assets like a site, a diagnostic, or a book, those reach people through your channels, and that visibility opens doors to speaking, clients, and partnerships. Three moves this week: connect the work you already have instead of making more of it, give people one clear next step so your visibility actually leads somewhere, and run your content through a repeatable system so it does not depend on willpower. The throughline is simple: you are not short on expertise, you are short on the system that makes it visible.

Editorial

I have spent a good part of June building a complete authority system for an expert who had everything except visibility. He had a documentary, a Ted talk, years of research, and almost no one outside his field knew any of it existed. The fix was not more content. It was connecting the work he had already done. The key point here is about building a system that multiplies the foundation you already built.

On June 26 Anthropic reported in its Economic Index “Cadences” survey : access to AI is now near-universal, so the edge is no longer having tools or even having expertise. The edge is a system that turns what you know into what people see, trust, and buy.

In this issue I walk you through the playbook the helps you turn your experience into authority. This extends the June 29th AI Blueprint: Knowing Isn't Enough: Build the System That Makes You Visible. If you are moving from a role into your own thing, this issue names the first asset to connect instead of the tenth post to write. If you already run a business and want one move this week, this issue shows you where authority leaks into nothing and how to close the gap. If you sell your judgment for a living, this issue names why depth of expertise no longer sells itself, and what does.

When AI access is universal and expertise no longer sells itself, the edge is a system that turns what you know into what people see, trust, and buy.

Juan

Move 1: Map your flywheel. Draw your core idea in the center and the four parts around it: source authority, public assets, channels, opportunity paths. Mark what you already own. Most experts are surprised how much is sitting idle.

Move 2: Find the one broken link. It is usually a fuzzy core idea, a funnel with no next step, or a content system running on willpower. Name the single weakest connection, not all of them.

Move 3: Add the front door. Stand up one diagnostic or clear next step that lets a buyer test your idea against their situation. Authority without a door converts into nothing.

Move 4: Build the memory file. Write one markdown file with your positioning, offer, and voice, and load it at the start of your next three content sessions. It is the cheapest upgrade to output you can make.

Try my second brain here and build your memory file

Run them in order and each move sets up the next.

Here is my experience to authority program, shoot me an email or book a 15-min free assessment when you visit the web: learn.buildtothrive.co/build-authority to learn more.

Clarity Prompts

Try these three prompts to help you narrow your focus and understand where your authority system is leaking and how to improve it.

The Authority Flywheel Audit. Walks away with a map of the four parts of your authority system and the one connection that is missing or weak. Best for the professional deciding where to start. (Open the prompt in the Prompt Library:

GO TO PROMPT

The Front-Door Namer. Walks away with one diagnostic or next step that turns visibility into a real conversation, priced as an engagement not a favor. Best for the operator whose authority converts into nothing. (Open the prompt in the Prompt Library:

GO TO PROMPT

The Voice Memory File. Walks away with a one-page markdown file your AI reads first, so drafts sound like you in one pass. Best for the operator whose content system runs on willpower. (Open the prompt in the Prompt Library:

GO TO PROMPT

Run together, the three prompts are the week’s playbook: map the system, add the door, build the engine.

Note: If you like to work solo, my AI Offer Build Coach helps you turn your expertise into a valuable offer, click on the banner to go to the tool. It is offered at $67 dollars, but it is free as part of your paid subscription which is now running at a 25% discount- Only $60 dollars for all prompts. Offer valid until July 31st.

Hey!— Juan here. This is a weekly read from Build to Thrive for operators and professionals turning hard-won experience into leverage, income, and opportunity in the AI economy. More than 5000 of you read along each week.

A word on why I write it. I have started four companies, and I am building my fifth: this one, a media company I run almost entirely solo, on a fleet of about thirty AI agents that draft this newsletter, scan my inboxes, prep my client work, and write my morning brief. I am not reporting on this shift from the sidelines. I am living it, and I share what works and what breaks, openly, as I go.

If you want to stop reading about it and build the system yourself, check out my Thrive Business System program. Two hours of one-to-one work with me plus twelve months of the Vault (more than 200 prompts, skills and workflows) and the premium Blueprint, where we turn your experience into a positioned, repeatable offer and a business that runs without you. No theory and no recycled frameworks. The same moves I run on my own fleet and with my clients, handed to you to use this week.

Thank you for reading Build to Thrive.

Subscribe now

A word on why I write it. I have started four companies, and I am building my fifth: this one, a media company I run almost entirely solo, on a fleet of about thirty AI agents that draft this newsletter, scan my inboxes, prep my client work, and write my morning brief. I am not reporting on this shift from the sidelines. I am living it, and I share what works and what breaks, openly, as I go.

If you want to stop reading about it and build the system yourself, that is Founder 100. Two hours of one-to-one work with me plus twelve months of the Vault (more than 200 prompts, skills and workflows) and the premium Blueprint, where we turn your experience into a positioned, repeatable offer and a business that runs without you. No theory and no recycled frameworks. The same moves I run on my own fleet and with my clients, handed to you to use this week.

The “Deli Counter” Trap

At the start of this year, I decided to run my business with AI agents. I didn’t sketch a plan or write a manual; I simply opened my laptop and started ordering agents like a man at a deli counter. “I’ll take a research agent. Give me a newsletter writer. And let’s throw in a Chief of Staff to run the whole thing.”

By Tuesday, I had twelve of them. By Wednesday, I realized I hadn’t built a team—I had built a relay race where everyone runs at the same time and nobody is holding the baton. My writer was already halfway through a draft using research that hadn’t arrived yet. My Chief of Staff was confidently summarizing work that didn’t exist. My software bill was climbing, yet I was essentially paying for a smoke detector beeping in an empty house: reports were being generated every morning that no human had opened in months.

The realization was as beautiful as it was dumb: AI isn’t a tool you “buy”; it’s a worker you must “manage.” Hiring agents without a management framework doesn’t save time; it simply promotes you, unpaid, to a full-time micromanager of software.

You Didn’t Hire a Team, You Built the “Worst Office in America”

The reason my twelve-agent “office” failed was that I skipped every fundamental of human management. In a traditional company, you’d never expect a new hire to succeed without infrastructure. With AI, we focus entirely on the Action (the hands moving) while ignoring the Infrastructure (the management of the work).

To move from firefighting to true leverage, you have to translate these twelve traditional management elements into your AI workflows:

• Job Profile: I failed to define what each agent owned, leading to a “turf war” where two agents produced the same article in different voices.

• Reporting Line: Because I didn’t establish who answered to whom, I became the answer to every question.

• Order of Operations: I lacked a sequence, so agents started all at once instead of in a logical chain.

• Trigger: I relied on a clock rather than a business event; one agent fired 11 out of 19 mornings when there was no work to do.

• Inputs: I provided no “memory file,” forcing agents to start cold and improvise every time.

• SOP: Without step-by-step instructions, the agents hallucinated their own processes.

• Handoff: I had no clean way to pass data, so work frequently “hit the floor” between steps.

• Definition of Done: I let “finished” be a matter of the agent’s opinion rather than my spec.

• Quality Check: I lacked a review gate, allowing a made-up statistic to ship straight to the page.

• Metrics: I didn’t check the logs to see if the agents were actually working or just being “busy.”

• Escalation Path: My agents were programmed to guess rather than raise a hand when they got stuck.

• Stopping Point: Without an “off switch,” one agent ran past a budget cap on a task with no end.

Stop Hiring “Staplers That Talk”

Most AI “failures” happen because we treat agents as tasks rather than loops. When you only focus on the middle of a job—the “doing”—you strip away the context and quality control that makes a human employee valuable.

“It is like hiring that great employee and keeping only the part where her hands move. That is not an employee. That is a stapler that talks.”

If your workflow is just a prompt that generates text without a clean start, a review, or a finish, you don’t have an automated worker; you have an expensive autocomplete. The “doing” is actually the least important part of the cycle if the beginning and the end are broken.

The Anatomy of a Perfect AI Loop

For an agent to be actually useful, it must operate within a structure that mimics a high-performing employee. There are five parts to a functional cycle:

1. A Trigger: A real business event (a payment, a form submission, a quiet lead). A clock is a poor trigger; a business event is a signal.

2. Signal and Context: The “memory file” or second brain. The agent must read your standards, history, and the brief before lifting a finger.

3. The Action: The execution. This is the one part everyone already has.

4. An Eval Gate: This is Andrew Ng’s “reflection pattern” in plain clothes. It is a logic gate or a second agent reviewing the work before a human—or a customer—ever sees it.

5. A Stop Condition: A declared finish line or budget cap to ensure the task doesn’t run indefinitely.

The Sixth Step: The Compound Interest of AI

The first five steps make the work run. The sixth step—Learning—is what makes the work compound. This is the rarest stage of AI management and the most vital for long-term strategy.

In a true loop, the agent captures what worked and what didn’t during a run and updates its own brief for the next time. This transforms the AI from a repeating cost into a compounding asset.

“You can offload a task, you can even offload a whole job, but you can never offload your learning.” — Satya Nadella

The software tool you use is rented and gets cheaper every month, but the “sharpening” of the process lives in your own files. If you skip this, your AI will be exactly as good in month six as it was on day one. If you include it, you are building an asset that you actually own.

Share

The 60-Second “Vibe Check” for Your Workflows

To see if you are a manager or a babysitter, audit your current AI use against this checklist. If you are missing these parts, your process is exposed.

• [ ] Does it start on an event, or a vibe? If the answer is “whenever I remember to run it,” you are the bottleneck.

• [ ] Does it read a brief first? Does it pull your past decisions, or do you re-explain the task every Tuesday?

• [ ] Does it do the actual work? (The Action step.)

• [ ] Does anyone check it before it ships? If the customer is the first one to see the AI’s mistakes, you’re missing an Eval Gate.

• [ ] Does it know when to stop? Is there a budget or a finish line?

• [ ] Does it learn? Does the next run start sharper because the results of the last run were captured?

The Engineer’s Caveat: Not every one-off question needs a six-part loop. Use the simplest thing that works for quick tasks. But for the recurring workflows you depend on, one out of six isn’t “simple”—it’s a liability.

From Firefighting to True Delegation

True leverage comes from building the loop around the work so the machine can manage itself. According to Anthropic’s “Cadences” report, which surveyed nearly 9,700 people, over a third of respondents expect AI to handle most or nearly all of their work within a year. The users who were most optimistic about this future were those delegating whole tasks, not just asking one-off questions.

The path from “more agents” to “complete loops” is the only way to get your time back. The babysitting ends when the loop handles the management for you.

Pick one workflow. Find the part you skipped. Close it this week.

Juan

Build Assets - Create Freedom - Thrive on your Terms.

I am excited to share a new tool to help operators find the constraint holding their business back, and where their next move (AI, automation, workflow redesign, clear ownership, or better measurement) creates the most leverage. It works whether they have no AI yet, are experimenting, or already use it everywhere.

Take a test drive!

learn.buildtothrive.co/mytimeback

Want to build the whole system with help? Inside my Founder 100 program I can help you turn your pile of half-working agents into a self-running operation: two hours of one-to-one work to help you streamline your process. I help you close the loops around your own work and build a self-running 5-person business, solo.

Join 5000 plus subscribers looking to thrive!

Subscribe now

A word on why I write it. I have started four companies, and I am building my fifth: this one, a media company I run almost entirely solo, on a fleet of about thirty AI agents that draft this newsletter, scan my inboxes, prep my client work, and write my morning brief. I am not reporting on this shift from the sidelines. I am living it, and I share what works and what breaks, openly, as I go.

If you want to stop reading about it and build the system yourself, that is Founder 100. Two hours of one-to-one work with me plus twelve months of the Vault (more than 200 prompts, skills and workflows) and the premium Blueprint, where we turn your experience into a positioned, repeatable offer and a business that runs without you. No theory and no recycled frameworks. The same moves I run on my own fleet and with my clients, handed to you to use this week.

Editorial

If you are an expert who keeps being told to “just post more,” this issue is for you.

I am building a visibility and authority system for an experienced academic right now. Decades of real knowledge, and almost no audience for it. And the same pattern keeps showing up everywhere: deep expertise has stopped compounding on its own. What compounds is a system that turns what you know into what people see, trust, and eventually buy.

Knowing your field used to be the edge. Now it is the price of entry.

The hard part is no longer the knowledge. It is the machine around it: a clear position, a repeatable way to publish, channels that connect, and a path that gives the right people a next step.

This issue walks the three moves that build that machine:

• Position. Pick the one idea you want to be known for, before you write another word. Not sure what that is yet? Here is a free assessment I have been using with clients, Experience to Offer Assessment. it will help you turn your experience into a viable offer.

• Systematize. Trade random posts for one repeatable workflow that compounds. Grab the free Context Files cheat sheet, the five-file system I build with clients, so AI writes in your voice every time.

• Convert. Make sure the attention you earn has somewhere to go. A call to action. I connect people to 50 Prompts from 50 Creators, it is a gem. The most viral post ever written by me and not surprisingly the one with the most value. I give it away. However, if you want to navigate through all of them on one single index that allows you to pick the one depending on the occasion, I offered it to paid subscribers. Prompt Navigator

If you are a consultant or fractional, this is how expertise starts generating opportunities instead of just respect. If you run a business, it is how your knowledge becomes a channel you own. If you are repositioning, it is the gap between being qualified and being chosen.

Knowledge is the raw material. The system is the leverage. Pick the move that fits where you are this week.

Juan

Most of your audience is not ready to buy, and that is the opportunity. Chris Tottman, who spent 30 years scaling software companies, estimates only about 3 percent of a market is actively buying at any moment (The Founders Corner). The system you build for the other 97 percent is the edge.

Knowing the tools is not the same as a return. MIT’s NANDA initiative found 95 percent of enterprise AI pilots show no measurable P&L impact, because the tools never connect to a real workflow (Fortune).

The advantage is the system, not the model. Satya Nadella, Microsoft’s CEO, argues the winners will own the learning loop between their people and their tools, not whoever picks the best model (Yahoo Finance).

You can hold the tool and never run it. Raghav Mehra, who writes Cash and Cache, notes 420 million people have Copilot and almost none run it as a system (Cash and Cache).

Owning a channel beats adding headcount. Adam Robinson’s RB2B reached roughly 8 million dollars in revenue on five people and one owned channel (LinkedIn).

Michael Simmons, who writes Blockbuster Blueprint to more than 115,000 readers, argues that as AI floods the feed with average content, the standout work becomes rare and valuable again, “caviar,” and the differentiator is the human expert behind it (Blockbuster Blueprint).

What it means: this is good news for you. When everyone can generate competent content, the scarce thing is a clear, trusted point of view. Your position, the one idea you want to be known for, is the asset. Without it, more output just adds to the noise you are trying to rise above.

What I am observing: with my client, the first sessions were not about writing. They were about deciding what he should be known for. Everything got easier after that.

John Brewton, who writes the Operating by John Brewton, sells his coaching with one line that stuck with me: “Most coaching ends with notes. This ends with a machine”. Nadella’s point is the same at enterprise scale: own the system, not the tool.

What it means: a post you write once and forget is a note. A position, a repeatable workflow, and connected channels are a machine that keeps working while you sleep. The expert who builds the machine pulls away from the one publishing on impulse. It is the same loop we walked on June 22, pointed at your own authority this time.

What I am observing: Often times, people do not t have an idea problem, they have a system problem. In my practice I build the system, not more notes.

Tottman’s core argument is that most companies fight over the 3 percent already shopping and ignore the 97 percent who are not ready yet (The Founders Corner). The same is true for any expert launching anything.

What it means: visibility with no next step is applause, not income. Every piece needs one clear path: subscribe, assess, book, or join. Build the path before the launch, not after, or you launch into silence.

What I am observing: the funnel is the least glamorous part and the one that turns a following into a business. We built my client’s next step before his first post went out.

Move 1: Name the one idea you want to be known for. Before you write anything, finish this sentence: “I help [who] do [what].” That single line is your position, and it filters everything else.

If you want help on this, I created a 7-process system that take you there AI Offer Builder

Move 2: Build one repeatable workflow, not random posts. Decide where ideas get captured, drafted, and published, then run the same path every week. Consistency comes from the system, not willpower.

Move 3: Connect your channels so nothing dead-ends. Posts point to your Substack, Substack to your site, your site to one offer. A pile of accounts is not a presence.

Move 4: End everything with one next step. Every piece earns the right to ask for one thing. Pick it on purpose.

Run them in order. Position first, or the rest just amplifies the wrong thing.

Prompt 1: The One-Sentence Position. Walks away with the single idea you want to be known for, written as one clean line. Best for an expert who is everywhere and known for nothing.

GO TO VAULT

Prompt 2: The Content Workflow Map. Walks away with a repeatable weekly path from idea to published, so posting stops depending on motivation. Best for an operator who publishes in bursts.

GO TO VAULT

Prompt 3: The Dead-End Audit. Walks away with every place your content ends without a next step, and the one step to add. Best for someone with an audience but no funnel.

GO TO VAULT

Together the three prompts build the spine: decide the position, run the workflow, close the loop.

THIS WEEK AT BUILD TO THRIVE

Two ways to turn this issue into your move.

The fastest start is already above: the free assessment that turns your experience into a viable offer at learn.buildtothrive.co/valueoffer, and the free Context Files cheat sheet for briefing AI in your own voice. Both ship free to subscribers this week.

Want to build the whole system with help? Inside Founder 100, $180 (typically $499), we turn your expertise into a positioned, repeatable content engine: two hours of one-to-one work plus 12 months of the Vault and premium Blueprint. I help you turn your experience into income and build a self-running 5-person business, solo, at learn.buildtothrive.co/founder-100.

No hype, no fluff. Just the one link worth fixing first.

This week’s companion, You’re Not Short on Expertise. You’re Short on a System That Makes People See It., drops Wednesday. See you next Monday.

Turn your experience into income and build the self-running business AI now makes possible. Weekly, specific, no fluff.

Upgrade to paid subscriber so we connect directly via the chat.

Subscribe now

Read that the way Alibaba framed it and you get the optimistic headline: AI lowered the barrier so far that one person can now build what used to take a team (Alibaba.com via PR Newswire). That is real, and if you have been waiting for permission to turn your experience into your own thing, this is the clearest sign yet that the market moved in your direction.

But there is a second read, and it is the one that matters more if you are the one building. When the barrier drops for you, it drops for everyone. If 71 percent of the people in the room are solo and 89 percent are leaning on the same handful of plug-and-play AI tools, then starting a one-person business is no longer the rare move. It is the common one. The thing that used to set you apart, that you started at all, is now table stakes.

So the question quietly changed. It is not “can I start a one-person business with AI.” Thousands of people just answered that for you. The question is “what makes mine the one that is still here in two years, when the field is full of businesses that look almost exactly like it.”

The part the tool cannot give you

Here is the mechanism underneath the surge. A plug-and-play AI tool is, by design, the same for everyone who buys it. That is the whole point of plug-and-play. It hands a first-time founder the same industrial design, the same marketing copy, and the same workflow it hands the founder next to them. The tool is the great equalizer, which is exactly why it cannot be your edge. Anything the tool gives everyone cannot be the reason a buyer picks you.

What the tool cannot supply is the part you already have and most people do not: the specific judgment that comes from years in your seat. The pattern you recognize on a sales call that a generic model would miss. The thing you know not to do because you watched it fail once. The way you frame a problem that makes a buyer feel understood in the first five minutes.

This rule comes from two pieces I have written, and they are two halves of the same point. In How Anthropic Set Off a Trillion-Dollar Chain Reaction I traced how fast the AI model itself is turning into a cheap commodity, the same one anyone can rent for a few dollars. In The Death of the High Performer I made the other half of the case: once the tool does the task, being good at the task stops being the thing you get paid for, and your judgment becomes the scarce part. Put them together and you get the rule for this moment. The tool races toward commodity, so the value moves to the judgment you wrap around it, the part a competitor cannot rent. In a one-person business, that wrapper is you, made specific.

The economics, in plain terms

Think about what happens to price in a crowd of lookalikes. When a hundred solo operators offer the same AI-built service with the same tool, the only lever any of them has left to compete on is price. So price falls. That is not a moral failing, it is just what happens when buyers cannot tell two offers apart. The technical word is commoditization. The plain version is that sameness pushes price toward the floor.

Now picture the operator whose offer carries something the tool did not supply. A buyer who can name why they want that specific person is not comparing on price anymore, they are comparing on fit. That operator holds their number while the lookalikes race each other down. Same tool underneath, completely different economics on top. The difference is entirely in the layer the buyer can see that did not come out of the box.

A move you can make this week, free

You do not need a new tool for this. You need to get specific on the one thing the tool cannot copy, and put it where a buyer can see it.

Try this. Write down the last three times someone paid you, or paid a company you worked at, for a judgment call rather than a deliverable. Not the report, the call. The moment where your read on the situation changed what happened next. For each one, finish this sentence: “Most people in my field would have done X, I did Y, and here is why it worked.” Those three answers are the raw material of an offer a competitor running your same tool cannot reproduce, because they did not live the thing that taught you Y.

That is the difference between “I use AI to build marketing campaigns” and “I have run growth for two consumer brands, I know which of these AI-built campaigns will quietly burn budget, and I build the ones that do not.” Same tool. One of those sentences competes on price. The other one does not.

When you want the structured version

The free move above gets you started. Turning those three sentences into a one-sentence offer you can actually sell, with a price and a shape, is the work I do inside Founder 100. If you are an experienced operator stepping into your own thing, the Positioning path is built for exactly this: I help you take your raw experience and shape it into a one-sentence offer, a priced engagement, and the language to put it on your profile and in a proposal. Session two turns that into a 30-day plan with three moves that put it in front of buyers. In case you like to drive solo, you can also explore this yourself with a 7-sequence prompt that takes your there. Offer Builder

If you are reading this as one of the 71 percent, or about to be, the tool already did its job. It got you in the room. What keeps you in the room is the part of you the tool could never supply. Pick the one judgment call you are proudest of and start there.

Juan

Editorial

I have been mapping how operators get AI to run without them through, and one word keeps deciding who pulls it off: the loop. I had written about this last February as a shift in competitive advantage for business operators. (Stop Renting Labor: Why “Owning the AI Loop” Is the New Solo Competitive Advantage). The signal sharpened last week when Satya Nadella, Microsoft’s CEO, told a room of executives that the advantage in an AI economy will not come from picking the best model. It comes from owning the learning loop between your people and your systems, what he called your token capital (News9Live). The enterprise crowd heard validation. The solo operator should hear a warning: if the advantage goes to whoever owns the loop and you do not have one, you have a pile of prompts you run by hand. If you read the June 1 Anthropic chain-reaction Blueprint and The Death of the High Performer, this is the build that follows. In this ediiton, I will walk you through the three forces that turn a prompt into a loop.

If you are a professional sizing up your first system, this issue gives you the five parts of a loop so you can build one this week. If you are an operator running a business, it names the part of your work worth encoding so it becomes an asset no competitor can rent. If you are deep in delivery, it shows you where loops quietly run badly and how to catch it.

Force one: a loop, not a prompt. A prompt is something you type, wait for, and read. A loop starts without you, pulls what it needs, acts, checks itself, and stops. Most operators built only the acting part, the AI that drafts an email or a proposal. That is a power tool, not a loop. The five parts are the trigger, the context, the action, the eval gate, and the stop. Miss one and the system stalls or runs badly without you noticing.

Force two: own the loop, own the asset. Nadella’s point in plain terms: the model gets cheaper and everyone rents the same one, so the durable advantage is the loop you built from your own work (News9Live). McKinsey’s State of AI 2025 found the firms capturing real value are the ones redesigning whole workflows, not automating single tasks (McKinsey).

Force three: the gate is where it holds. The eval gate is your definition of good, applied before anything ships. It is where your judgment gets encoded, the part no competitor can buy from OpenAI or Anthropic.

The loop is the structure, owning it is the advantage, and the gate is where your standards live. Pick the one force that fits where you are this week.

Juan

Build to Thrive caters to ambitious professionals and founders reinventing themselves in the AI economy. It explores how people can use their experience, tools, and ideas to create leverage, income, and opportunity in a business world being rewritten.

Join 5000+ subscribers

Copy the infographic below and save it….this sums up the concept and structure of loops. Keep reading if you like to have the prompts, tools and tactical moves that help you build loops for your business.

What happened. Nadella named the shift in enterprise language: competitive advantage comes from controlling the learning loop between staff and systems, where people supply judgment and corrections and the system improves future work (News9Live). The operator version is more concrete: a system that handles the execution layer of a recurring task end to end.

What it means. You can build one this week if you map the five parts. The trigger starts it without you (a Monday 8 AM timer, a new intake form, a deal gone silent for five days). The context is what it reads first so it acts on your situation, not a blank guess. The action is the draft or the score. The eval gate checks it against your standard. The stop tells it when to quit. Build those five and the Monday morning that used to take 90 minutes of piecing things together takes 15 minutes of decisions.

What I am observing. What keeps landing for me is that most operators already have the action and call it AI. The other four parts are where the leverage hides, and they are the cheap part to build.

What happened. Nadella drew the line between two kinds of capital: human capital (your judgment and relationships) and token capital (the AI capability you build and own) (EdTech Innovation Hub). McKinsey’s read points the same way: whole-workflow redesign is what separates the firms capturing value (McKinsey).

What it means. The model is a rental everyone shares, so it cannot be your edge. The loop you built from your own clients, your own data, and your own standard is the thing that holds value, and the capacity it frees is what you sell. When the execution layer of a recurring task runs without you, you take one more client or buy back two hours for business development. That is the loop converting into revenue.

What I am observing. The way I read Nadella is that this is permission, not theory. The operator who owns one real loop owns something a competitor cannot rent from a model provider.

What happened. The reason people say AI almost works is the missing gate. An eval gate is a definition of good applied before anything ships: a human read, an automated check, or a second AI pass. Skip it and you ship the 80 percent and spend your hours fixing the last 20. The math is unforgiving: an 85 percent reliable step, run 10 times, lands near 20 percent end to end (Cash & Cache).

What it means. The gate is where your standard gets encoded, and the stop condition is what keeps the loop from sending its fourteenth email to a prospect who left nine emails ago. Build both and you stop babysitting the AI, because you only ever see output that already passed. Over time the gate gets tighter and the loop learns what good means in your specific business.

What I am observing. What I keep seeing is that the gate is the part operators want to skip and the part that decides whether they trust the system. It is also the part that is purely yours.

Move 1: Pick the most boring task in your week. Not the strategy or the client calls. The Monday pipeline review, the Friday status update, the daily lead check. The repetitive work that feels the same every time is where a loop pays off first.

Move 2: Map the five parts. Write down what triggers it, what context it needs, what it produces, what good enough to ship looks like, and when it stops. Those five answers are the spec for your first loop. You walk away with a build plan, not a vague intention.

Move 3: Build the eval gate before the action. Define the check that catches the 80 percent stuff before it reaches a client. This is the part that turns a first-draft machine into a system you trust.

Move 4: Set the stop condition out loud. Decide the rule that ends it (”three follow-ups, then mark cold and stop”). Writing the stop forces you to define success and failure, which is worth more than most plans in your drive.

The Loop Mapper. Walks away with one weekly task mapped to all five parts, trigger through stop, ready to build. Turns a vague “I should automate this” into a concrete spec. Open the prompt

The Token Capital Audit. Walks away with the part of your work worth encoding into a loop, the judgment a competitor cannot rent. Names the asset hiding inside your recurring work. Open the prompt

The Eval Gate Builder. Walks away with a written definition of good enough to ship plus a stop condition for one loop. Turns the reliability risk into a check the system runs before you see anything. Open the prompt

Run together, the three prompts are this week’s playbook: map the loop, name the asset, build the gate.

The tool that fits this week is Claude Code’s /goal command, which gives an agent a verifiable finish line so it stops when the work is actually done Claude Code /goal: Stop Babysitting Your AI Agent. That is the stop condition and the eval gate in tool form: the agent keeps going until a check says the goal is met, then quits.

You do not have to use the command to take the lesson. Before you build any loop, write the finish line a stranger could check. That one habit is what separates a loop that runs from one you have to babysit.

If you want to build your first loop with help, Founder 100 is $99 for two sessions of 1:1 coaching plus 12 months of the Vault and premium newsletter, at learn.buildtothrive.co/founder-100. Session one names the constraint and builds the loop. Session two pressure-tests it after two weeks in the field.

That’s all for this edition, thank you for reading!

…and remember.

Build assets. Create freedom. Thrive on your terms.

Hello, I am Juan. I have started four companies, and two 7-figure exits. I am building my fifth: a media company for operators and professionals chasing leverage, income, and opportunity in the AI economy. I help businesses grow, and I share what I learn as I go, openly, with you. More of my work, free resources and how I can help is at juansalasromer.com. I hope this article earns your time.

Subscribe now

One morning, concerned about my token costs, I discovered that my briefing agent had read about twenty files, opened four full transcripts of the day before, and paged through every contact in my address book, all to hand me a half-page summary of my day. Picture a human assistant doing that: photocopying the entire filing cabinet every morning to write you one sticky note. You would catch it in a week. Because it was an agent, the work looked instant and the cost stayed invisible. That is the whole trap.

In May 2026 a line from AI engineer Andrej Karpathy spread among builders: roughly 90 percent of a typical AI coding bill pays for context the user did not need to send. The shape is easy to picture. An AI tool loads 50 files into memory to answer a question about 30 lines of code, and you pay for all 50, used or not. Run that fifty times a day and the unused reading, not the answers, is most of the bill.

Here is the way I think about it now. Your AI is not charging you to think. It is charging you to read. Every agent I run is a fast, eager assistant who reads everything you put in front of it before it answers, and bills you for the reading whether it used a word of it or not. You are probably not writing code. I am not either. I run Build to Thrive on a fleet of about fourteen of these assistants that draft my newsletter, scan my inboxes, prep my client work, and write my morning briefing. Different work, identical bill. The mistake is the same on both sides: you go looking for a cheaper assistant when the real leak is the stack of paper you keep handing the one you already have.

This is the same lever I wrote about in this week’s Blueprint, You Do Not Need a Better AI. You Need a Better Brief. The AI Blueprint | Week of June 15th, from the other side. A vague brief is expensive twice. You pay for the rough draft you have to fix, and you pay for the pile of context you dumped in hoping the model would sort it out. Cut the pile and the bill follows, with the output unchanged.

WHAT YOU WILL NEED

Before you start, you need very little:

- One AI tool or agent you run on real work, the more often the better.

- The one workflow you run most. Frequency is where waste compounds, so that is where you look first.

- A way to see what your tool loads before it answers. If you cannot see a token meter, and most of us cannot, you will audit by counting instead, which the next section covers.

- About two hours, and a willingness to read your own logs.

THE ONE PROMPT THAT FINDS THE LEAK

You do not need a token meter to find waste. Take the one workflow you run most, then paste this into your AI before you change anything:

Whatever it flags as loaded-but-unused is your bill, leaking. The reread-every-time pile is the standing waste you fix once. The five steps below are what you do with that list.

THE STEPS

Step 1: Start with your worst offender

Why it matters. Waste compounds with frequency, so the cheapest win is hiding in the thing you run most, not the thing that feels most expensive.

What to do. Point the prompt above at your highest-frequency workflow first. Read what it loads against what it actually used.

Example. I started with the agent that writes my morning briefing, because I run it five days a week. It was reading about twenty files a run, pulling up to four full transcripts of the prior day’s work sessions, and opening my entire contacts folder whether or not anyone in it came up. Most of it never reached the brief.

Pitfall. Audit the thing that runs, not the copy you assume is real. My first move was to optimize a file that turned out to be a stale duplicate of the instructions, not the version actually running. I almost spent an hour tuning a file that does nothing.

Step 2: Send the part, not the pile

Why it matters. Most “I need the whole file” instincts are wrong. The model needs the section that matters, not the whole document, and you pay for the difference every run. Hand your assistant the one folder the answer is in, not the whole cabinet.

What to do. For anything stable that rarely changes, stop sending it every time. Send the relevant part, or send it only when it has actually changed.

Example. Three of my twenty files were locked documents that almost never change: my positioning, my business brief, my org map. The agent re-read all three daily. Now it reads them only when they change, and the session pull dropped from thirty messages to fifteen because the brief only ever uses the tail of a conversation.

Pitfall. The “just in case” include. You add the file because you are not sure the model needs it. Default to leaving it out and letting the model ask.

Step 3: Stop the rereading

Why it matters. A long session re-reads its entire history on every new turn, and a standing context file re-reads the parts that never move. Both are pure waste.

What to do. Close long sessions with a short handoff and open a fresh one. Keep your standing context in one file, and make the tool stop re-reading the stable parts of it.

Example. I closed a bloated work session with a five-line handoff and started clean. That is the same move, run on the work itself.

Pitfall. Reaching for prompt caching as the fix. Everyone recommends it, and it barely helped me. Caching is a thermos: it keeps your context warm for the next pour, a gift if you are pouring cups all afternoon and useless if your next cup is tomorrow morning. My agents run once a day, so the thermos is always cold by the time they wake up. Caching pays off when you fire many times in a short window. For a daily schedule, the lever is not caching what you load, it is loading less.

Step 4: Cut the cadence where frequency was the cost

Why it matters. Some agents are not expensive per run, they are expensive because they run too often for what they return.

What to do. Find the heavy jobs running more often than the work requires, and space them out. Make sure they pick up where they left off so nothing is missed.

Example. Two of my agents ran a browser scan every day. I moved the two heavy ones to every other day. They stop at the first item they have already seen, so they just clear a slightly bigger batch. That one change removed about eight heavy runs a week.

Pitfall. Spacing out a job that does need daily cadence. Cut frequency only where a longer gap changes nothing about the result.

Step 5: Decide what does not need a brain

Why it matters. The biggest lever coders talk about is routing mechanical work to a cheaper model. On a no-code fleet you often cannot assign a model per task, but you can ask a sharper question: does this job need a thinking model at all? You do not book a surgeon to put on a band-aid, and you do not need a reasoning model to move a file.

What to do. For purely mechanical jobs, copying a file, moving a record, stop running an AI for them and move them to a plain script.

Example. I pulled several mechanical steps out of my agents entirely. You cannot always route a task to a cheaper brain. You can often decide it does not need a brain.

Pitfall. Scripting something that actually needs judgment. Keep the model on anything where the right answer changes with context.

THE FINISHED RESULT

Here is the before and after on my worst offender, with no change to what it produces.

Before. The briefing agent read about twenty files every run, pulled up to four full session transcripts, and opened my entire contacts folder, five mornings a week. Two other agents ran heavy daily scans. Earlier in the year, before any of this, three more agents were doing work already done elsewhere.

After. The agent reads only what changed, pulls the tail of a session instead of the whole thing, and opens a contact file only when that person comes up. The heavy scans run every other day. Switching off the redundant agents in the spring had already cut my daily spend by roughly half. The briefing reads exactly the same. The bill does not. It stopped photocopying the cabinet, and nobody on the receiving end can tell.

The input-side version of this same discipline is the brief. This week’s Blueprint ships a free one-page Delegation Brief Build to Thrive | The AI Blueprint | Week of June 15th, 2026, the four-part template for specifying work so the model reads less and guesses less on the way in. The audit prompt finds what your agents over-read; the Delegation Brief stops you over-feeding them in the first place. They are two ends of one habit.

If you would rather see where else your operating model leaks time and money before you touch a single workflow, the free AI Leverage Assessment (https://learn.buildtothrive.co/mytimeback) runs that read in about twelve minutes. If you would like to turn this into a priced, margin-disciplined service build it out inside Founder 100 (https://learn.buildtothrive.co/founder-100). I am running a special promo for $99 dollas for a 2-hour session.

It is dull work. It is also the line between a service that scales and one that just keeps you busy.

Juan

50 Creators Shared Their Rock-Star Prompts. Here They Are

Six months ago, I started reaching out to creators I was curious about, plus a few I was lucky enough to get referred to. Mostly one virtual coffee at a time.

The ask was simple: would you be open to sharing your rock-star prompt with the @Substack community in my Monday AI Blueprint edition, and have your newsletter featured alongside it?

Nothing here was scraped or borrowed. Every prompt was shared deliberately and generously, by people who had no reason to hand over their sharpest work — except that they believe knowledge is worth more when it moves.

And move it did. The depth, the ingenuity, the sheer practicality of these creators left me humbled. This edition is a small monument to their generosity.

So today, with a lot of gratitude, I’m sharing a consolidated volume of 50 prompts from 50 creators. Thank you to every one of you who said yes. I hope you all enjoy it.

Join 4500+ subscribers

Subscribe now

EDITORIAL

The post that got operators talking last week was not about a new model. A growth and operations operator, Jordan Murphy, promoted a tool on one promise: it does not hand you back a rough draft you then spend two hours fixing (LinkedIn, week of June 12 - Tier 3, market observation). It struck a nerve, and the replies were a wall of operators saying the same thing. After eighteen months of the “70 percent draft,” they do not want a co-pilot anymore. They want a finisher. And almost nobody is telling them how to get one.

Here is the part I keep coming back to. The reason your AI hands you a rough draft is almost never the model. It is the brief. Two operators on the identical model get wildly different output, and the only variable between them is how they specified the work. Tsedal Neeley, a Harvard Business School professor who studies how companies actually absorb new technology, has made the same point: sharp context produces better output at a fraction of the cost (Tier 1, voice; exact-quote verification owed). The model layer is commoditizing in front of us, Opus 4.8 and Anthropic’s trillion-dollar filing are the same story, so the durable, ownable skill is not the AI. It is what you wrap around it. This week, that wrap-around is the brief.

If you are a professional building your own thing, this issue hands you the one skill that turns AI from a research assistant into an actual employee. If you are an operator handing off delivery work, it is the difference between AI that saves you time and AI that quietly makes more of it. The May 4 Stack Audit (https://www.buildtothrive.co/p/build-to-thrive-the-ai-blueprint-3e5) and the May 11 Money Stack (https://www.buildtothrive.co/p/the-money-stack) walked the tools and the pricing. This issue walks the skill that makes the tools pay off.

You already know how to do this. When you hand work to a sharp new hire, you give them context, tell them what good looks like, and show them an example. You skip all of it with AI because the little box invites a one-liner, and a one-liner gets you an average draft no matter how good the model is.

So pick one deliverable you run every week, write it a real brief this week, and see what comes back. The free one-page Delegation Brief at the bottom is the template.

Juan

Join 4500+ subscribers

Five numbers behind why the brief, not the model, is your constraint.

Two hours. The time operators keep losing to cleaning up an AI draft that should have come back finished, which is the exact pain a real brief removes. It is the complaint under every “AI saves me time” caveat: the draft arrives at 70 percent and the last 30 eats the afternoon (operator observation - Tier 4).

~90%. Andrej Karpathy, a well-known AI engineer, estimated that about 90 percent of a typical AI bill pays for context the model never needed (via Ronin, mid-May 2026 - Tier 3). A sloppy brief is not just lower quality. It is also where the waste lives, because you pay for everything you dump in whether the model uses it or not.

$1.50 and $9.00 per million tokens. Gemini 3.5 Flash runs there (Google Gemini pricing - Tier 2: https://ai.google.dev/pricing). When the model is this cheap, the output quality is almost entirely in your hands, not the price tier. The brief is the lever, not the spend.

Close to a trillion dollars. Anthropic filed to go public at a $965B valuation while the model keeps getting cheaper (Fortune, June 1, 2026 - Tier 1: https://fortune.com). The model is becoming a commodity. The skill you wrap around it is the part that stays yours.

Most jobs, not all. Briefed well, today’s AI finishes a clear majority of tasks but not every one, a sharp jump from a year ago (Stanford AI Index 2026 - Tier 1, primary-source confirmation owed before publish). That gap is why a one-minute spot-check on the predictable failure points stays part of the job.

Subscribe now

Three signals, one skill: brief the AI like a staffer, not a search box.

Story 1: The Rough Draft Is a Briefing Failure, Not a Model Failure

What happened. Jordan Murphy’s post hit a nerve because it named the thing operators are tired of: the draft that should have been finished but comes back needing two hours of cleanup (LinkedIn, June 12 - Tier 3, market observation). The reaction was not about the tool. It was operators recognizing their own daily frustration in one line.

What it means. Give a model a thin instruction and it fills the gaps with generic, average assumptions, because that is the safest guess. A new hire in the same spot would stop and ask you three questions first. The model does not ask. It guesses, and it guesses average. So a one-line prompt guarantees an average draft no matter how capable the model is. The quality you are missing is not locked inside a better AI. It is sitting in the four things you did not tell it.

What I am observing. The operators getting finished work are not on a secret model. They write the AI a real brief, the same way they would brief a person they respect. The skill is plain and teachable, and almost nobody is teaching it, which is exactly why the demand has nowhere to go.

Story 2: Even Briefed Well, Spot-Check It

What happened. The same operators raving about AI finishers are quietly still checking the work. Briefed well, today’s agents finish most tasks cleanly, a large jump from a year ago, but not all of them (Stanford AI Index 2026 - Tier 1, primary-source confirmation owed). The successful teams keep a quick human check in the loop while they learn where the workflow breaks.

What it means. A good brief gets you finished work. A quick spot-check keeps it honest. The failures are predictable, which is what makes them fast to catch: invented specifics, facts that were true once and are now stale, and numbers that do not hold. You do not re-read the whole thing. You check those three places, the way a manager skims a new hire’s first few deliverables, and then you ship. Speed without a check is how a good workflow ships a bad fact.

What I am observing. The operators who trust their AI most are the ones who still spot-check it, because the check is what earns the trust. Drop it and you are one fabricated source away from an embarrassing send.

Story 3: A Tighter Brief Costs Less, Too

What happened. Karpathy’s line that about 90 percent of an AI bill pays for context the model never needed spread among builders for a reason (via Ronin, mid-May 2026 - Tier 3). With Gemini 3.5 Flash at $1.50 and $9.00 per million tokens (https://ai.google.dev/pricing), the model is cheap, so the waste is almost entirely in what you send it.

What it means. A vague brief is expensive twice. You pay for the rough draft you have to fix, and you pay for the pile of context you dumped in hoping the model would sort it out. A tight brief is the opposite: it tells the model exactly what to use, so it reads less, guesses less, and bills less. If you sell AI-leveraged work at a fixed fee, the brief quietly sets your margin. This is the money beat under the skill, not the headline, but it is real.

What I am observing. The operators running the best margins treat the brief as a cost control as well as a quality control. Same habit, two payoffs.

The four parts of a real brief. You already give all four to a person. Give them to the AI and the draft comes back finished.

1. Context. What a smart new hire would need to know about the situation, the audience, and the goal. Not the whole company history, the part that changes the work.

2. The deliverable, defined. What exactly should come back: the format, the length, the structure. “A competitive analysis” is a wish. “A one-page table comparing five competitors on price, positioning, and one weakness each” is a brief.

3. The standard. What good looks like here, and the two or three things that would make it unusable. This is the part everyone skips, and it is where average output hides.

4. One example. A reference, a sample, or a past piece at the level you want. One good example does more than a paragraph of description.

Before and after, on a real deliverable. Before: “Write me a go-to-market plan for my new service.” Back comes a generic plan built for an average business that is not yours. After: the four parts, your buyer, your constraints, your price band, and one plan you admired as the reference. Back comes something you can edit in ten minutes instead of rebuild in two hours. Five minutes to brief beats two hours of cleanup on a deliverable you run every week.

________________________________________

Three prompts that build the briefing habit, so you leave with the skill, not just the idea. Full versions in the Notion Prompt Library, linked in each card.

The Delegation Brief Builder. It interviews you for the four parts, context, deliverable, standard, and one example, then assembles a clean brief you can paste into any model and flags the part you left thinnest. Walks you out with a brief tight enough that a stranger could run the task. GO TO PROMPT

The Before-and-After Rewrite. Paste your one-line prompt and it shows you the average draft that prompt is quietly asking for, names the three missing inputs, and rewrites it as a real brief side by side. The fastest way to feel why the one-liner fails.

GO TO PROMPT

The Manager’s Spot-Check. Paste an AI deliverable and it runs a manager-style review of the three places AI work breaks, invented specifics, stale facts, and shaky numbers, and gives you a ship-or-not verdict. Keeps the speed without shipping a mistake.

GO TO PROMPT

Run in order, the three are the delegation loop: write the brief, see why it works, check the result. Management, not prompt tricks.

Claude Projects (https://claude.ai) is the tool this week, and ChatGPT Projects work the same way. It is where you write your brief once and reuse it on every task.

What it is, plainly: a Project is a workspace that holds a standing set of instructions plus a few reference files. Everything you do inside it starts already knowing your business, your voice, and what good looks like, so you stop briefing from scratch every session. It is the four-part brief, made permanent.

How to set one up, in about five minutes:

1. Create a new Project and name it after a deliverable you run often, like “Weekly Competitive Scan.”

2. Fill out the Standing Brief context file once (the free template below): who you are, your audience, your voice and standards, your default format, and one example.

3. Paste it into the Project’s instructions, then upload one or two reference files, a past piece you were proud of, a style note, a data source you reuse.

4. From then on your prompt can be a single line, because the brief already lives in the Project.

5. When the output drifts, fix the context file once instead of re-explaining yourself every time.

Grab the fill-in-the-blanks Standing Brief context file, spend ten minutes on it, and you have turned the briefing skill from a per-task chore into an asset that compounds. You are not writing better prompts. You are building a better-managed employee whose onboarding doc you only write once.

For the thesis underneath this week, read James Presbitero The Only Moat Left In The AI Age - by James Presbitero. His argument is that the only moat left in the AI age is the thing a competitor cannot copy or prompt their way to. The brief is the operator-scale version of exactly that. The model is the commodity everyone shares. How you direct it is the part that is yours, and it is the part that finishes the work.

Two ways to turn this issue into your move.

Grab the free one-page Delegation Brief. It is the four-part template from this issue on a single page, the thing to keep next to your keyboard until briefing the AI like a staffer is automatic. It ships free to subscribers this week. While that is going out, the fastest way to see where your own operating model is leaking is the free Diagnostic, a roughly 12-minute, question-asking assessment at learn.buildtothrive.co/valueoffer (https://learn.buildtothrive.co/valueoffer).

Want to build the skill out fully? Inside Founder 100, $99, typically $360, the briefing discipline is where the “Your Head of Content” and Chief of Staff work begins: two hours of one-to-one with me to turn your recurring deliverables into briefed, finished AI workflows. I help you turn your experience into income and build a self-running 5-person business, solo. learn.buildtothrive.co/founder-100 (https://learn.buildtothrive.co/founder-100).

Join 4500+ subscribers: https://buildtothrive.co/subscribe

50 Creators Shared Their Rock-Star Prompts. Here They Are

Six months ago, I started reaching out to creators I was curious about, plus a few I was lucky enough to get referred to. Mostly one virtual coffee at a time.

The ask was simple: would you be open to sharing your rock-star prompt with the @Substack community in my Monday AI Blueprint edition, and have your newsletter featured alongside it?

Nothing here was scraped or borrowed. Every prompt was shared deliberately and generously, by people who had no reason to hand over their sharpest work — except that they believe knowledge is worth more when it moves.

And move it did. The depth, the ingenuity, the sheer practicality of these creators left me humbled. This edition is a small monument to their generosity.

So today, with a lot of gratitude, I’m sharing a consolidated volume of 50 prompts from 50 creators. Thank you to every one of you who said yes. I hope you all enjoy it.

A word on why I write it. I have started four companies, and I am building my fifth: this one, a media company I run almost entirely solo, on a fleet of about thirty AI agents that draft this newsletter, scan my inboxes, prep my client work, and write my morning brief. I am not reporting on this shift from the sidelines. I am living it, and I share what works and what breaks, openly, as I go.

Want the right tool for your business, not just another list? Book a 15-minute call. I will assess your bottleneck and point you to the tool or system that fits. If it needs more digging, I will come back with the answer on a follow-up call

15-min tool delivery

Welcome to the first edition of the B2T Vault: 50 Essential Prompts for Operators.

Every prompt here came from a creator I met over the last six months across , and almost all of them over a virtual coffee. I reached out to each one personally. Nothing here was scraped or borrowed. It was shared, deliberately and generously, by people who had no reason to hand over their sharpest work except that they believe knowledge is worth more when it moves.

Six months in, here is what I keep coming back to. Substack is a standing invitation into other people’s hard-won experience. Open a publication and you often find someone who has already solved the exact problem in front of you, and is glad to show you how. The depth, the ingenuity, the sheer practicality of these operators is humbling, and this edition is a small monument to it.

This spirit is not unique to these pages. A few weeks ago I was invited to join , a community built by creators who believe the same thing this edition is built on: that knowledge compounds when it is shared. It was founded by , , and , and two of them, Joel and Claudia, also appear in this Vault. That tells you how small and generous this corner of the internet really is. Getting the invitation felt less like joining an organization and more like being welcomed into a conversation that was already underway.

To every creator in these pages: thank you. This Vault is yours as much as it is ours. And to anyone reading who builds in public and has a prompt worth sharing, consider this an open door. Vol. II starts now.

How the Vault is organized

A flat list of fifty prompts is a pile, not a tool. So we organized them around the work itself, the recurring jobs an operator hands to AI over a week:

• Leadership and Decision-Making: making the call you already have the information to make

• Content and Storytelling: shaping a piece of writing that proves the point

• Product and Building: shipping the minimum version of the thing

• Growth and Marketing: finding and moving the next customer

• Operations and Systems: turning a one-off into a repeatable process

• Research and Synthesis: making sense of what you read

• Career and Personal Brand: naming your value in a sentence a buyer can act on

• Creativity and Craft: stress-testing your own thinking

• Mindset and Operating: getting your head right to do the work at all

Those nine cover the operating surface of a one-person business. Whatever is in front of you today, it almost certainly lives in one of them. The payoff is speed to the right prompt: you go to the cluster that matches the problem on your desk and find a prompt a working operator already built, used, and trusted for exactly that. You are not borrowing a clever idea. You are borrowing a tested move.

Read it slowly. Each entry is a door, not a download. Below every prompt you will find the person who wrote it and the newsletter they publish, because the prompt is only the surface. The real value is the operator behind it and the work they put out every week. Take your time. Open the ones that catch you, read the description, and follow the people whose thinking you want more of.

The range alone is worth the wander. You will meet , a poet and professor who asks you to pause before you prompt. , a former Navy helicopter pilot who treats AI adoption like a preflight checklist. Christopher Ljams, a security researcher who writes as and shows you how AI systems actually break. , who reads the space economy for a living. , an investor mapping which startups break out and why. , an educator charting AI’s real effect on the classroom. Fifty operators, fifty bodies of work, one afternoon well spent.

Enjoy the journey.

Juan

The prompts in this cluster share a single operating assumption: that the most expensive thing a leader does is delay a decision they already have the information to make. Each of these was built by someone who manages without a management layer, which means they could not afford to confuse process with progress. What they built instead are tools for getting to a clear call faster, with less noise in the room.

01 · The Founder’s Pre-Mortem For pressure-testing a big bet before you make it. Assume the business already failed, then let a security-style post-mortem expose your single points of failure. by ToxSec (Christopher Ljams) · ToxSec → Read the prompt

02 · VC-Grade Pitch & Report Analyzer For reading a pitch or report the way an investor would. Strips the hype and returns an investment-readiness snapshot. by · The Founder-VC Vibe → Read the prompt

03 · Protocol for Structured Thinking For high-stakes or ambiguous decisions. Forces three sharp clarifying questions before any answer, so you decide on evidence, not a confident guess. by · AI + Education = Simplified → Read the prompt

04 · The Theory of Change For when a big goal feels abstract. Works it backward into a causal chain that ends in the actions to take this week. by · Startup Riders → Read the prompt

05 · AI Adoption Obstacle Prompt For when an AI rollout keeps stalling. Diagnoses the people, process, and structural blockers and hands you a five-move starter plan. by · Getting AI to Work → Read the prompt

Share

Every creator in this cluster figured out the same non-obvious thing: the story is not the vehicle for the idea. The story is the proof. These prompts are not about making content more engaging. They are about making an argument the reader cannot dismiss, because the argument arrives wearing the coat of someone’s lived experience. The prompts here are structural. They give the idea a skeleton before they give it a voice.

06 · Article-to-Post Prompt For turning a long article into LinkedIn and Twitter posts that travel, built around the reader’s interest, not your story. by · Write Your Way To Wealth → Read the prompt

07 · Content Repurposing Prompt For getting more from one piece. Turns a single blog article into a LinkedIn post and a four-minute YouTube script. by · Aisha’s Notes → Read the prompt

08 · Substack Note Editor For writing a Substack note that cuts through. Distills a topic to four sentences with a hook, often a contrarian one. by · The Circuit → Read the prompt

09 · The Technical-to-Human Copy Translator For when your product is technical but your buyer is not. Rewrites architecture into copy about outcomes, with the jargon stripped out. by · Vibe Stack Lab → Read the prompt

10 · Cross Domain Lens Generator For when your topic feels commoditized. Borrows principles from an unrelated field to produce angles a skeptical buyer respects. by · Robots Ate My Homework → Read the prompt

11 · The Disruptive Keynote Speaker For building a talk that challenges a room. Three provocative, credible points that reframe an industry problem as a choice. by · Substack → Read the prompt

Find where your business can gain time, increase income, and create happier customers.

AI Leverage Assessment

The builders in this cluster share a bias toward constraint. Every prompt here was written by someone who did not have an engineering team, a design sprint, or a six-month runway. They had a problem, a deadline, and a clear picture of what done looked like. The prompts reflect that. They are fast tools for people who cannot afford to prototype their way to an answer. Build the minimum version of the thing. Ship it. Learn from contact with reality.

12 · Digital Product Opportunity Generator For deciding what to build. Surfaces scalable product opportunities your audience will actually pay for, ranked by revenue. by · The Digital Creator → Read the prompt

13 · Emotional UX & Product Reality Check For when a screen works but feels off. Maps the emotional journey and ties each fix to a user emotion and a business metric. by · Eidos Design → Read the prompt

14 · PRD Generation Prompt For going from idea to a clickable app. Produces a spec an AI prototyping tool like Lovable or Replit can build from, with no dead links. by · UX + AI → Read the prompt

15 · Product Backlog Impact For deciding what to build next. Ranks five backlog features by impact, revenue, complexity, and strategic fit. by · Product Release Notes → Read the prompt

16 · Production Code Checklist For shipping AI-written code you can trust in production. A checklist the AI must clear for security, memory, and failure modes before it writes. by · Wondering About AI → Read the prompt

17 · The AI Fluency Partner For experts who want to build, not just use AI. Scaffolds the path from idea to a working product. by · Build to Launch → Read the prompt

Growth for solo operators is not the same problem as growth for a funded startup. The constraint is not capital, it is attention. The prompts here were built by people who had to figure out how to move the right person from curious to committed without a sales team, a CRM pipeline, or a retargeting budget. They are tools for making the right reader feel specifically seen. Specificity, in every one of these, is the mechanism.

18 · The Money-Making System Prompt For when you have skills and an audience but no system to monetize them. Maps your hidden assets into three products you could launch in under ten hours. by · AI Made Simple → Read the prompt

19 · Account Signal Hunter For researching a target account before you reach out. Separates real buying signals from noise and turns the strongest into outreach. by · The Lucky Seller → Read the prompt

20 · PR Agency Prompt For getting a read on anyone’s online reputation. A digital PR snapshot plus an 80/20 strategy and a 30/60/90-day plan. by · How to AI → Read the prompt

21 · Offer Revamp Prompt For when your offer is fuzzy. Turns any offer or member benefit into one sharp, outcome-driven sentence from the reader’s point of view. by · Unplugged → Read the prompt

22 · AI Brand Integrity System For keeping AI-generated content on-brand. Locks your voice, philosophy, and visuals before you hit generate. by · AI Flight Plan → Read the prompt

23 · Competitive Intelligence Prompt For understanding a market fast. An evidence-based competitor analysis with a positioning map and where to differentiate. by · One Million Goal → Read the prompt

This article gives you free access to all 50 prompts. You can access our prompt navigator to easily use the prompts based on your particular need, plus access to our entire inventory and premium content by becoming a paid subscriber. We are running a 25% special discount. Offer available until August 15th. ($60/year)

25% discount

These are the prompts built by operators who were tired of doing the same thing twice. Not automators, systems thinkers. Each of them, at some point, sat down and decided to convert a recurring judgment call into a documented process. The prompts here are the tools they used to do that conversion. They are not about efficiency. They are about making decisions durable, so that what you figured out once does not disappear when you close the laptop.

24 · Oryn AI Mini For turning one bottleneck into a system. Five questions in, you get a named workflow and a tiny Make.com automation. by · Cove Connect → Read the prompt

25 · The What’s Actually Broken? Scan For when the business feels chaotic but you can’t name why. Pinpoints your three biggest friction points and the one fix to make this month. by · Systems and Sideyes → Read the prompt

26 · The Context Portability Prompt For carrying a long AI conversation into a fresh thread. Exports it as a clean Markdown handoff with usage instructions. by · Product Strategy Decoded → Read the prompt

27 · The System Failure Pre-Mortem For an AI or data system you depend on. A five-point audit that finds the hidden point of failure before it finds you. by · The Data Letter → Read the prompt

28 · The Trust Protocol Shift For evaluating or redesigning an AI security system you actually need to trust. A seven-step audit for transparency and accountability. by · Lights On → Read the prompt

The bottleneck in most operator research is not access to information. It is the gap between reading something and knowing what to do with it. The prompts here were built by people who read widely and get paid to know what it means. They are translation tools, from raw signal to actionable intelligence. Each one compresses a reading-to-decision cycle that used to take hours into something a well-briefed AI can complete in minutes. The operator still makes the call. The prompt does the compression.

29 · Audience Intelligence Research Prompt For getting past assumptions about your audience. Researches and structures real audience intelligence into a usable brief. by · AI Blew My Mind → Read the prompt

30 · SpaceTech Intelligence & Investment Lens For evaluating an emerging-tech company or market. Systems-level analysis grounded in real economics, not hype. by · Cyclop SpaceTech → Read the prompt

31 · The Content Analyst Prompt For analysis you can trust. Makes the AI reason step by step and ground claims in evidence to cut hallucinations. by · Cash & Cache → Read the prompt

32 · Your Content Research Co-Pilot For finding fresh angles your audience hasn’t heard. A Claude Project that knows your Substack and surfaces contrarian takes. by · Wander Wealth → Read the prompt

33 · 3x3 Content Matrix Analyzer For making sense of what you read. Classifies any article into a 3x3 matrix of intent and scope you can reason from. by · Automato → Read the prompt

Share

These prompts were built by people who went through the uncomfortable work of naming what they do in a sentence that a buyer could act on. Not an elevator pitch. Not a LinkedIn headline written for a recruiter. A sentence that holds up when someone says “can you help me with X” and you either know immediately that you can, or you know immediately that you cannot. The prompts here are the tools they used to get there, and the ones they now use to help clients do the same.

34 · Find My Ikigai For when you’re rethinking your direction. Uses the Ikigai framework to find your path across passion, skill, market, and impact. by · Solopreneur Code → Read the prompt

35 · Skills-to-Income Strategic Action Plan For turning what you know into income. A monetization plan and 30-day roadmap across freelancing, coaching, and products. by · Learn Grow Monetize → Read the prompt

36 · Voice Architecture Analysis Prompt For making AI write like you. Builds a voice spec from your samples that captures how you actually think on the page. by · Silicon and Soul → Read the prompt

37 · The Anti-Mission Coaching Prompt For when goals feel vague. Defines your anti-mission, what you refuse to tolerate, for sharper clarity than chasing “success.” by · Purplemind → Read the prompt

38 · Context-Aware AI Engine For getting AI off generic defaults. Aligns it to your ICP, business model, and voice so output reflects your strategy. by Julia Díez · The AI-Ready Localizer → Read the prompt

39 · The CMO Concept Distillation Prompt For explaining a complex concept to an executive fast. Turns it into a clear briefing using Pareto and the Feynman technique. by · Uncharted → Read the prompt

Subscribe now

The operators in this cluster treat AI as a sparring partner, not a ghostwriter. They are not using it to generate content they would not have written. They are using it to stress-test ideas, find the better version of an argument, and catch the assumption they are too close to see. These prompts have a different texture than the others in the Vault, they are designed to produce friction, not to remove it. The friction is the value.

40 · Strategic Sparring Partner For stress-testing a plan instead of being agreed with. Turns AI into a partner that challenges your logic and sharpens direction. by · The AI Maker → Read the prompt

41 · How to Beat AI Confirmation Bias For when AI is just agreeing with you. Paste it after any answer to force a real devil’s advocate and reality check. by · Leadership in Change → Read the prompt

42 · Critical Claim Reviewer For pressure-testing an idea or offer. Flags claims that won’t hold up and rewrites each into a more defensible version. by · What Just Happened in AI → Read the prompt

43 · The AI Prompt Evaluator For improving a prompt you already use. Returns three optimized variations that cut tokens and add clear guardrails. by · Product with Attitude → Read the prompt

44 · The Cognitive Shield (Quantitative Mode) For telling human writing from AI. Scores whether a text is AI-generated, AI-assisted, or fully human, with a read or do-not-read verdict. · The Airlock → Read the prompt

45 · AI Verification Coach For learning to trust AI output without being fooled. Coaches you through recognize, verify, refine, and apply. by Tawnya Means · The Collaboration Chronicle → Read the prompt

This cluster is the hardest one to explain and the easiest one to recognize. The prompts here are not about a deliverable. They are about the state of mind the operator needs to be in to produce good work, and how to get there faster, with less internal resistance. These were built by people who have run their own thing long enough to know that clarity is a discipline, not a mood. The prompts here are the tools they use to manufacture it.

46 · AI Reflective Partner For using AI with intention, not speed. Gentle reflective prompts to reclaim attention and meaning. by · Slow AI → Read the prompt

47 · The AI Execution Momentum Builder For when you’re overloaded and stuck. Three questions in, it names the single next step that reduces the most friction. by ClariSynth · ClariSynth → Read the prompt

48 · Guided Year Retrospective For closing out the year with honesty. Walks you through a facilitated reflection, then writes it back in your own voice. by · Behind the Screens → Read the prompt

49 · The Micro-Map Planning Prompt For overwhelm that isn’t really laziness. An ADHD-aware coach that closes the extra tabs and builds a seven-day path you can follow. by · Business Feels Good → Read the prompt

50 · The Noise Filter For cutting through a swirl of inputs. Slows the conversation and names the real decision in front of you. by · whatisthat.ai → Read the prompt

Closing

These fifty operators-built tools sharp enough to feature and were generous enough to share them. If you would like to be featured in a future edition of our weekly AI Blueprint, send me an email to juan@buildtothrive.co

Want to be able to quickly navigate through all these prompts? Join the Vault for a complete inventory of prompts and workflows, already indexed and easy to navigate when you have the specific need.

Prompt Navigator