9 Comments
User's avatar
Mike Goitein's avatar

The hardest piece, but the greatest value-add to me, is the quote from McKinsey:

"McKinsey’s State of AI 2025 found the firms capturing real value are the ones redesigning whole workflows, not automating single tasks"

It's lazy to automate existing processes. Way harder to reimagine what is only now just becoming possible.

Juan Salas-Romer's avatar

That’s a game changer. Absolutely

Dr. Michael Meneghini's avatar

Strong framing of AI value shifting from prompts to systems; durable advantage comes from owning workflows, evaluation, and feedback loops, not just model access.

Juan Salas-Romer's avatar

Thank Michael. Appreciate your feedback

Daniel Ionescu's avatar

The stop condition may be the most useful part here.

Al will happily keep going long after the work has stopped being useful.

Very relatable, to be fair. 😅

Juan Salas-Romer's avatar

Totally. We all learn this the hard way! 😂

Dennis Berry's avatar

Shifting focus from writing better prompts to designing closed loops, especially with an encoded "eval gate" is the exact moment an operator stops playing with AI tools and starts building actual enterprise equity.

Cory Blumenfeld's avatar

i used to think the gate had to be automated to count

turns out the gate that works starts as just you reading every output. you hand pieces to the ai as you trust them, and thats when you stop babysitting

John Brewton's avatar

Most operators built the action and called it AI. The other four parts are where the leverage actually lives.