01 The Problem

Most people who think they're building with AI are still prompting.

They've learned to write better prompts. Maybe they've chained a few together and called it a workflow. What they've actually done is get faster at asking the same questions from scratch every day.

The gap between prompting and building isn't technical. It's conceptual. Prompting is the floor — the level playing field that everyone with a ChatGPT account can reach. Building is something different: persistent context, systems that run, an operating rhythm designed around how you actually think.

Most people will stay at the floor. The gap is widening every month.

02 What It Is

A 3-week sprint. 12 people max.

One weekly Zoom call, daily async support in a shared channel, and a 30-day check-in after the cohort ends.

You come in with your real work. You leave with an AI operating system built around how you actually think — not a framework you'll forget by the following Monday.

Not a course. Not a bootcamp. A working sprint with a small group of people taking the same problem seriously.

03 Who It's For

For people who've started — and feel like they're circling.

You're already using AI daily. You've been at it for at least 6 months. You're getting output — sometimes good output — but you still feel like you're starting from scratch every time you open a chat window.

You've noticed something else, too. The thinking that used to happen in the friction is happening less. The work is easier. That should feel like progress.

It doesn't quite.

BTP is not for people who haven't started yet. It's for people who've started and feel like they're circling.

04 What Happens

Three weeks of live work. Built on a Week 0 you do alone.

Week 0

00

Delivered async
before the cohort starts

The Audit

Before anything else, you map what you're actually doing. Not what you think you're doing. A structured self-assessment covering which tools you use and how often, what you're genuinely offloading versus what you still own, and where your thinking is engaging versus coasting.

The audit creates the gap the cohort closes. Without it, every framework that follows feels generic. With it, everything clicks against real behavior.

Week 1 — Tuesday, June 9

01

90-min live session
+ async channel

Prompts to Workflows

The session opens with a live exercise. The same question, two containers. Once in a chat window where the answer evaporates. Once written to a file on your filesystem — a document you own, can edit, can build on. The debrief: what changed? Not the information. The relationship with the output.

Then: the Four Principles. Before any framework, the philosophy. Think before prompting. Simplify to the minimum. Make surgical changes. Define what done looks like before you start. These four explain why most prompting is still prompting even when people believe it's something more.

Main teaching: the 3-layer workflow model. Context layer — what the AI needs to know before you ask. Execution layer — the difference between a prompt and a brief. Review layer — how to evaluate output without outsourcing your judgment.

Deliverable

One core workflow rebuilt from scratch using the 3-layer model.

Week 2 — Tuesday, June 16

02

90-min live session
+ async channel

Operating Rhythm

The question isn't just how you work with AI. It's when.

Deep work blocks stay AI-free. That's where the cognitive reps happen — the thinking that has to remain yours. Synthesis blocks bring AI in after you've done the work. Recovery is non-negotiable.

You'll design your Personal Performance OS: a one-page document describing your daily rhythm, where AI fits in, and where it doesn't.

You'll also build your first recurring skill: the Weekly Intelligence Briefing. Every Friday, the agent pulls what you need to know across your specific domains and writes a structured report to a file. Not a chat answer that disappears. A document waiting for you when you sit down.

Deliverables

Personal Performance OS · Weekly Intelligence Briefing running.

Week 3 — Tuesday, June 23

03

90-min live session
+ diagnostic prep

Diagnostic

Content delivery ends in Week 2. Week 3 is about what actually happened.

Before the live call, you complete a six-question diagnostic. Which skills ran every day. Which ran once. Which never ran. Where the output felt off. Where it surprised you. Michael reviews responses before the call. The session runs on this foundation, not on first-pass discovery.

Then: Skill 10. The answer to the highest-friction moment you named in the diagnostic. Not a template. Not a pre-built script. You name the gap, write a one-paragraph spec, and build it live. By the end of the session, it's running.

Deliverables

System snapshot · Skill 10 built and running · one forward commitment said aloud to the group.

+30 Days

30

One call
One month after Week 3

The 30-Day Check-In

One call, a month out. What held. What broke. What surprised you. An accountability anchor — and a chance to reconnect with 11 other people who built the same thing.

05 What You Leave With

Seven things. All of them yours.

01

A personal AI audit

Baseline picture of how you were actually using AI before — the honest version, not the assumed one.

02

A redesigned workflow

One core process rebuilt with intentional 3-layer structure. Yours to use, modify, and extend.

03

A Personal Performance OS

One-page daily rhythm document integrating AI without sacrificing the deep work that requires only you.

04

A Weekly Intelligence Briefing

A configured, running skill that pulls what you need to know every Friday and writes it to a file. Automatic. Waiting.

05

A system snapshot

The OS as it actually runs at the end of Week 3 — not as designed, but as it held up against a real week of work.

06

Skill 10

Built from your own diagnostic gap. Not a template. Yours.

07

Eleven other high-agency people

The smallest useful cohort. Everyone in the room at the same level, taking the same problem seriously.

06 Who's Teaching

// portrait — Michael Puhala

Michael Puhala teaches what he uses — not what he read about.

Michael is VP of GTM and Product Strategy at Bevy. He's been building AI workflows inside an actual operating role since 2023 — not in demos, not in a lab, but in the daily work of running a GTM motion at a real company.

He's the author of Human Traits and the founder of The Drop-In newsletter. BTP runs the same way he does: everything in the curriculum comes from what he's built and what he's tested. The cohort is the proof.

Role

VP, GTM & Product Strategy — Bevy

Author

Human Traits

Founder

The Drop-In newsletter

08 Pricing & Logistics

$250

Cohort 2 · One-time

Seats

12 maximumHard cap. No exceptions.

First call

Tuesday, June 9 · 4:00 PM CDTThen weekly through June 23

Format

Zoom · async channel · 30-day check-in

Refund

Full refund if you cancel before Week 1No questions asked

Secure your spot — $250

09 Questions

The honest answers.

Is this recorded?+

Sessions are recorded for cohort members only. The live interaction is the point — recordings are a backup, not the product.

What tools do I need?+

Whatever you're already using. The Week 0 audit meets you where you are. The cohort is tool-agnostic.

Is this right for me if I'm just starting with AI?+

Probably not. BTP is for people 6 months or more in who still feel like they're circling. If you're earlier in the journey, start with The Drop-In.

What's the refund policy?+

Full refund if you cancel before Week 1 begins.

Move from AI prompting to autonomous agentic workflows.

Building, not prompting

A cohort of 12

Your real work