Here’s that call on paper: three statements, written before the build, in the team’s own words. These are BobCo’s, a made-up company with a very real problem:
- STOP
- If the assistant needs customer data we can’t send to the model’s vendor, we scrap it and put the people somewhere else.
- PIVOT
- If the model bill grows faster than the tickets it closes, we cut the cost before we scale, not after.
- INVEST
- If support agents keep using it after the pilot ends, without being asked, we make time to talk about expanding it.
BobCo’s assistant is made up. My agents aren’t. They’re internal tools I built on open-source models to make building cheaper: a nutrition assistant, deck builders, copy editors. I build with Claude too. That’s 15 years of doing the job I teach.
I teach at MICA, where I designed and teach the Product Design and UX Capstone. At Columbia, I designed and taught Technology Management and AI/ML in 2023–2024. I’ve taught my hands-on AI workshop for a corporate client through First Principles, the training partner that brought me in. I’ve also run master classes for IxDF, a live session for O’Reilly and a 320-person workshop at Mind the Product.
Building is the easy part now. A product manager can get a working prototype out of an AI before lunch. So teams pile up half-built things faster than ever, and nobody’s sure which ones deserve to live. Who decides? Usually the loudest person in the meeting.
That’s why I teach the call on paper, like BobCo’s, written before anyone falls in love with the prototype.
Three entries, ready to forward
You’ll have to explain me to someone who’s never met me. Pick the entry that fits your room and send it as is.
A hands-on AI workshop for product managers on the call AI makes urgent: what to stop, pivot or invest in once building is cheap. In the 90-minute version, learners build an AI assistant, then write the stop, pivot and invest statements for it. They need a laptop and a free Claude account, and they leave with the assistant and a written call on its future.
It also runs as a half day, online or in person.
Taught by Adam Thomas, Head of Product at Flax & Teal, who has built internal agents on open-source models (a nutrition assistant, deck builders, copy editors) and builds with Claude. He has taught for First Principles, IxDF and O’Reilly.
I teach product managers to use AI well. I’ll teach your curriculum or bring mine.
Hands-on AI workshop for product managers · 90 minutes
For: product managers, including people who have never used Claude.
Bring: a laptop and a free Claude account.
What you’ll do: start with worked examples and a quick warm-up build, like a web page, so newcomers catch up. Then the main build: an AI assistant, followed by its stop, pivot and invest statements, the conditions that would kill it, change it, or earn it more time.
You leave with: an assistant you built and a written call on its future.
Led by: Adam Thomas, Head of Product at Flax & Teal. He ran a 320-person workshop at Mind the Product and teaches product at MICA.
Also available as a half day.
Checking me as a conference workshop lead? I’ve run a 320-person workshop at Mind the Product. I also gave a breakout at #mtpcon Digital 2020, “The Initiative Document”. It wasn’t about AI.
Adam Thomas · university teaching
MICA: Part-Time Faculty, Product Management MPS (current). Designed and teaches the Product Design and UX Capstone.
Columbia: Lecturer, M.S. Technology Management (2023–2024). Designed and taught Technology Management and AI/ML.
At both schools: full 16-week terms, with sections as large as 70. He wrote the syllabi, graded and ran the LMS himself.
Degree: BS in Computer Science. Both schools appointed him, with that degree and 15 years in industry, to design courses for their master’s programs. He is Head of Product at Flax & Teal now.
Available: adjunct appointments from the next term, online or on campus.
If anyone asks whether I can hold a room, send this with the entry. It’s from an online webinar I hosted in 2024. The subject is metrics, not AI. Watch the loop: I explain one idea, set a 5-minute exercise, then come back and push the room on what they wrote in chat.
My hands-on AI workshop has run for a corporate client through First Principles but has no feedback to show yet. Mike took an earlier session of mine on the stop, pivot and invest call:
I’ve been using the approach to talk about a particular pilot I’m engaging in with a start-up. I’ve been very clear that there is a difference between a pilot and an enterprise roll-out and that we need to determine what are the non-negotiable success criteria for our organization to make sure that we are fully aligned on whether we kill, pivot, or move forward.
Mike
What we’ll settle first
The format
A conference slot, a partner class and a term-long course need different things. I’ll ask which one you’ve got.
Online or in person
I teach both ways. Tell me which you need.
This class is wrong for a room that wants a tour of AI tools, a demo to watch, or someone to tell them what to build. I don’t do transformation programs, I don’t do interim leadership, and I don’t consult. I teach.
One more thing, up front: the call only sticks where people are allowed to make it. Learners who can’t say stop at work will leave with a skill they can’t use yet. I’d rather you hear that from me now than from your feedback forms.
Bring me in to teach.
For a partner program, a conference workshop or a university course: tell me who’s in the room and when.