From 'Run a Workshop' to AI Builder Sprint: What Happens When AI Pushes Back on Your Plan
From “Run a Workshop” to AI Builder Sprint: What Happens When AI Pushes Back on Your Plan
Meta Description: We came in with a simple 2-hour workshop idea. Claude pushed back, ran the revenue math, and helped us design something that could do $87k year one. Here’s that conversation.
The original plan was simple: design a 2-hour AI Business Accelerator Starter Kit, find 5 warm prospects, deliver it, call it the first session.
That plan was fine. It wasn’t wrong. It would have gotten us to first revenue.
But it wasn’t the right plan — and the planning session with Claude Code is what surfaced why.
Where the Original Plan Fell Short
The “deliver a 2-hour workshop” plan had one problem: it didn’t compound.
You run a 2-hour session, collect feedback, run it again. Each session is a standalone event. There’s no community, no shared momentum, no reason for participants to talk to each other or to you between sessions. The content you create for session 1 doesn’t build toward anything beyond session 2.
We walked into the session thinking the question was “when do we run the first session?” Claude Code reframed it: “What kind of product are you actually building?”
The Insight That Changed the Direction
The reframe came from a simple observation about the ICP.
The primary target for ABT is the aspiring founder with a day job. They have an idea. They can’t execute it fast enough. They know AI matters but don’t have a system for it.
The specific failure mode we kept seeing in this audience: they think like a product manager OR they think like an engineer. Almost never both at the same time.
If you can think like a PM (define the problem, understand the user, scope the work) AND build like an engineer (ship something, iterate fast, close the feedback loop) — you can move as fast as a funded startup.
AI closes the gap between those two modes. One person can now operate both sides of the loop.
That’s not a workshop. That’s a transformation. And transformations take longer than 2 hours.
How the AI Builder Sprint Concept Emerged
The expanded concept came from the session itself — not from a prepared plan.
The direction came in pieces:
- 60-day cycle (2-week funnel, 30-day workshop, 2-week post-event)
- Solo builder track and team of 2 track — matching PM-brained with builder-brained participants
- Weekly live sessions: kickoff, office hours, show & tell, Demo Day
- Contest format with real prizes funded from registration revenue
- Scholarship spots per cohort to drive advocacy and social proof
- Email prework sequence to arrive on Day 1 ready to build
Each addition was tested against the 10-15 hours/week constraint. When the original 60-day workshop idea came up, Claude Code flagged it directly: weekly live sessions over 60 days is 8-10 hours of active delivery, which consumes most of the weekly budget. Recommendation: keep the workshop at 30 days, use the other 30 days for pre-event funnel and post-event follow-up.
That’s the kind of feedback a good PM gives. It doesn’t say “no.” It says “here’s how you get what you want inside the constraint you have.”
The Revenue Math That Made It a Decision
Before a single ticket was created for the Sprint, we ran the numbers across three scenarios.
| Run | Solo seats | Team seats | Gross | Prize fund (20%) | Net |
|---|---|---|---|---|---|
| Run 1 (pilot, launch price) | 20 × $297 | 5 × $497 | $8,425 | $1,685 | $6,740 |
| Run 2 | 30 × $397 | 10 × $697 | $18,880 | $3,776 | $15,104 |
| Run 3 | 50 × $497 | 15 × $897 | $38,305 | $7,661 | $30,644 |
| Year 1 (5 runs) | ~$130k | ~$104k net |
Run 3 alone clears the $10k/month exit number.
The prize fund mechanic is intentional. Giving back 20% of gross as contest prizes creates a real incentive for participants to promote — they are competing for money funded by their own registration fees plus everyone else’s. That’s a structural reason to share, not just a social ask.
Pricing is list-high, launch-discounted. Solo: $597 list, $297 early bird (first 20 seats). Team of 2: $997 list, $497 early bird. The discount creates genuine urgency for cohort 1 without permanently anchoring a low price.
What the Session Also Produced
The Sprint isn’t just a product. It’s the content engine for everything else.
- Weekly live recordings → course modules for AI-1001 and AI-2001
- Participant builds → real case studies, not hypotheticals
- Show & tell sessions → testimonials with context
- Demo Day → the kind of launch moment that generates video content, press, and social proof
Every hour of workshop delivery produces multiple reusable assets. That’s leverage — which is the only way to build a business at 10-15 hours per week.
The Sprint Structure We Landed On
60-day cycle:
- Days 1-14: Pre-event — signup open, email prework sequence, Discord setup, team matching
- Days 15-44: The 30-day workshop
- Days 45-60: Post-event — case studies, cohort 2 waitlist opens, recordings published
Inside the 30 days:
| Week | Theme | Live session |
|---|---|---|
| 1 | Think like a PM — problem, user, scope | Kickoff 90min |
| 2 | Build like an engineer — AI-assisted dev, iteration | Office hours 60min |
| 3 | Ship and validate — feedback, pivots, polish | Show & Tell 60min |
| 4 | Demo Day — final pitches, judge panel, awards | Showcase 90min |
Four live sessions. Roughly 8-10 hours of active delivery per cohort including prep. Sustainable at the constraint.
Program Identity
Name: AI Builder Sprint Tagline: Zero to Shipped Positioning: Learn while building. A community of builders and founders.
The name came last. The concept came first. That’s usually the right order — name the thing you built, don’t build the thing you named.
Launch Date: May 5
Signup page live May 5. Two-week pre-event. Workshop kicks off May 19. Showcase on June 13. Post-event wraps June 27.
Everything needed before May 5 is now a tracked ticket in Linear. The critical path is 4 weeks: landing page, launch funnel, prospect outreach, contest rules, scholarship application.
The session that produced this plan took about 90 minutes. Most of that time was thinking, not typing. The thinking was better because we had a structured partner — not just a tool — asking the right questions.
This is part 3 of a 3-part series on using Claude Code as a product manager:
- Part 1: The CLEAR prompt that structured the entire session
- Part 2: How we wired Claude Code to Linear’s GraphQL API
- Part 3: This post — the product thinking that emerged
AI Builder Sprint cohort 1 opens May 5 at alienbraintrust.ai.
Comments