You don’t break into AI Product Management by waiting for someone to hand you an “AI Product Manager” title. You get there by doing AI work now, using the tools you already have, shipping small projects, and telling compelling stories about the problems you’ve solved.

Plus the hiring market hasn’t caught up yet:

  • Companies write fantasy job descriptions (“5+ years AI Product Manager”, “enterprise AI observability experience”) for skills that barely exist.
  • Middle product management layers have been hollowed out.
  • Most AI Product roles get filled through networks and narrative, not job portals.

So the path looks less like climbing a ladder and more like creating undeniable evidence that you’re already operating as an AI Product Manager.

This is the playbook I’ve been following (and coaching others on).

Start With a Real Problem You Care About

My AI work didn’t start with, “I want to be an AI Product Manager.”

It started with, “I have a problem I don’t know how to solve without AI.”

For example:

  • A client needed to categorize retail products across messy marketplace taxonomies. Static rules and manual mapping weren’t scaling. That constraint pushed me to build an AI Product Categorization API that reads product attributes and selects the best category leaf node.
  • I was struggling to summarize my own career for different roles, so I built a Resume Bot that lets people chat with my experience instead of reading a PDF.
  • As a golfer, I wanted a better way to practice like I play, random shots, varied distances, mental reset between swings. That turned into a golf practice app concept that I used AI and vibe-coding to build that generates and track randomized shot plans.

None of these started as “AI projects.” They were just real problems where AI turned out to be the best lever.

That’s your starting point too: pick a real problem in your world that feels just beyond reach with ordinary tools, and lean into AI as the unlock.

Learn by Building, Not Browsing

If you’re reading this, you probably already have access to tools like ChatGPT, Claude, Cursor, Replit, v0, etc…

The difference between “I use these tools” and “I’m an AI Product Manager” is what you do with them.

When I built my Resume ChatBot, I:

  • Started with someone else’s rough idea.
  • Tried one stack, hit a wall, deleted everything.
  • Tried again with another tool, hit a different wall, deleted everything again.
  • Iterated until I had something I was comfortable putting my name on.

Same with my categorization API: it began as a scrappy experiment, evolved into a real Flask-based API with OpenAPI specs, and now it’s something I can walk a CTO through at the architecture level.

You won’t read your way into AI Product Manager. You’ll build, break, and rebuild your way into it.

Treat Prompt Engineering as a Product Superpower

Prompt engineering is no longer a party trick, it’s part of the Product Manager toolkit, like writing user stories or PRDs.

Good prompts are just structured thinking: clear context, constraints, tasks, and examples. The better you get at that, the better your AI tools perform across everything you do: research, ideation, UX copy, prototypes, even code.

👉 This deserves its own deep dive, so I’ve broken my framework out separately.

For now, the takeaway is simple: every min you spend getting better at prompts will pay off across every AI tool you touch.

Turn Your Work Into a Portfolio of AI Problems Solved

The projects I listed above aren’t just “cool demos.” They each anchor a story about AI in a product context:

  • AI Product Categorization API

    Problem: Retailers and marketplace sellers drown in SKUs that need to be categorized correctly across different taxonomies. Traditional rule-based systems are brittle and slow to maintain.

    What I built: An AI-driven API that takes product attributes + marketplace taxonomies and returns a best-fit leaf category, plus the full category path for human verification. Now I can talk through training data, evaluation, failure modes, and how it plugs into an existing catalog workflow.

  • Resume Bot

    Problem: Hiring managers don’t have time to truly understand a candidate’s depth, and I needed a way to demonstrate my AI + API chops in a single artifact.

    What I built: A chatbot trained on my experience that can answer questions like “How has Tim led API governance?” or “What’s an example of a product he took from 0→1?”. It’s both a portfolio piece and a live proof that I know how to scope, design, and ship an AI-powered experience.

  • Golf Practice App (in progress)

    Problem: Driving-range practice rarely looks like on-course golf. Unless you are Roy McAvoy, you’d never hit the same shot over and over on the course.

    What I scoped: A mobile-friendly app that randomizes shot distances and tracks thumbs-up/down feedback to reveal which distances I’m actually bad at. It’s a sandbox for testing AI around personalization, decision support, and progress tracking.

These aren’t theoretical. They’re things I can demo and deconstruct in an interview or a stakeholder conversation.

Why I Was Considered for an AI Product Manager Role (When I Applied for an API Product Manager Role)

Recently, I applied for an API Platform Product Manager role.

The first half of the interview was standard: I walked through my background, API platform work, governance, developer experience. It was fine, but not electric.

Then the hiring manager asked, “What have you been working on lately?”

I started talking about:

  • The AI categorization API: why I built it, how it sits in a larger retail catalog flow, where AI fits in the loop.
  • How I used AI tools to help design and implement it (OpenAPI first, Flask API, evals, n8n integration).
  • The Resume Bot and how I’m using it as a living artifact of my work.

He stopped me mid-stream and said (I’m paraphrasing): “The first part of the conversation felt like you were going through the motions. When you started talking about this AI project, your energy completely changed. I actually want to open an AI Product role. Would you be interested in that instead?”

Same comp band. Different problem. Same interview.

He saw an AI Product Manager not because my title said “AI Product Manager,” but because my stories and artifacts made it obvious that I was already doing the job.

(For the record, I didn’t get the role, rejected for lack of “enterprise AI experience,” which is another yet to be written blog post, but the important part is: the portfolio and narrative had me reframed as an AI Product Manager candidate.)

That’s the bar you’re aiming for.

A 30-Day Challenge to Kickstart Your AI Product Manager Transition

If you want to move from “interested in AI Product Manager” to “operating like an AI Product Manager,” here’s a concrete 30-day challenge:

  • Week 1: Pick a real problem.

    Choose something from your world, your team, your side gig, your hobby, that AI might help with. Write a one-page description of the problem and who it matters to.

  • Week 2: Draft a simple PRD and prototype.

    Use ChatPRD to outline the user, workflow, and success metrics. Then build the smallest possible prototype: a script, a basic chatbot, a single workflow. Done is better than impressive.

  • Week 3: Refine with better prompts and feedback.

    Iterate on your prompts. Ask a few friends or colleagues to “use” your prototype and talk out loud. Listen for where they’re confused or delighted and adjust accordingly. See my prompt framework here.

  • Week 4: Ship the story publicly.

    Write a short post or record a 2–3 minute Loom: What problem you tackled, what you built, where AI added value, and what you’d do next. Then share it. Tag people who might care.

When you finish, I’d love to see what you built.

Share your project with me, or hit me up if you’ve got an idea but don’t know where to start. I’m always happy to riff on use cases, scope something with you, or point you at patterns that have worked for me.