Here's the thing: I've spent the last few months building products with AI as my actual partner. Not "AI-assisted." Not "I asked ChatGPT for help once." I mean daily, intensive, building-real-things collaboration.
And I have thoughts.
This is not "another tool"
Let me be clear about what's happening. AI isn't just a fancier search engine or writing companion. It's changing how work actually happens.
I've shipped more in the last two months than I did in some entire years at big companies. Not because I'm working harder — because the shape of work changed.
The old loop was: think → align → build → review → ship.
Now it's more like: explore + draft + test + revise all at once. Everything happens in parallel. You can prototype something in an afternoon that would've taken weeks to spec out.
It's honestly kind of unhinged. In a good way.
What I've learned building this way
Judgment is the new bottleneck. When drafting is instant, choosing becomes the hard part. I went from "how do I get this done" to "which of these ten options is right." That's a fundamentally different job.
Discovery and delivery blur together. I don't write detailed requirements for simple builds anymore. Why would I? Just build the thing, see if it works, iterate. The prototype IS the spec.
The core job stays the same. Here's what hasn't changed: you still need to put the user at the center. You still need to solve real problems. You still need taste and judgment. AI just compresses everything around those fundamentals.
The frameworks might be dead
Two-week sprints. Quarterly roadmaps. Jira tickets with acceptance criteria.
I'm not sure these make sense anymore.
That doesn't mean I've stopped planning — I've just had to build entirely new systems. Skills documents that give AI context about projects. Our own ways of tracking KPIs and OGSPs. Session sprints focused on one thing at a time. A parking lot for ideas that aren't ready yet. Living docs that update as we go instead of specs that get stale the moment they're written.
I've been experimenting with something I call session sprints. One session, one focus, ship or learn something. AI and I move fast. We can pull data, do an analysis, write a blog, and build it into my website all in a day.
It's a whole new way of working. And honestly? It's working great for me as a solo builder collaborating with AI.
What I'm genuinely curious about is how this translates to enterprise. When you have 50 people who need to coordinate, does this approach scale? Do you need different systems? I don't have answers yet — but I know the old frameworks weren't built for this speed.
The old model was: decide → execute.
The new model is: explore continuously → decide deliberately → execute briefly.
How this could work in bigger organizations
Here's my hypothesis for how AI fits into real product teams.
Product uses AI to reduce uncertainty faster. Problem framing, decision prep, prioritization, communication. And to Draft PRDs and success metrics. Use it to explore risks and tradeoffs. PMs come to discussions with options, not questions. Less time explaining, more time deciding.
Engineering uses AI to collapse execution time. Faster implementation, earlier feedback, better collaboration with Product. Quality shifts left. Cycles compress.
The real shift is how they work together. Before: PM defines → Eng implements → Learn later. Now: PM and Eng explore together with AI. Multiple options surface early. Decisions happen faster. AI becomes the shared sandbox where ideas get tested and tradeoffs get debated. Capacity increases.
What AI doesn't do
Let me be clear about the limits.
AI doesn't own decisions. It doesn't set strategy. It doesn't replace taste or judgment. It's not accountable when things go wrong.
AI proposes. People choose.
The teams that get this wrong are the ones who think AI means less thinking. It actually means more thinking — just faster, with better inputs, and focused on the stuff that actually matters.
Where to start if you're a PM
Here's where AI actually helps in day-to-day product work:
Requirements and specs. Draft PRDs, user stories, acceptance criteria in minutes instead of hours. AI won't know your context perfectly, but it gives you a starting point to edit rather than a blank page to fill.
Fast mockups before full builds. Before you spend engineering cycles, have AI prototype the idea. See if the flow makes sense. Catch the obvious problems early. The prototype becomes the conversation starter.
Stress testing ideas. Ask AI to poke holes in your strategy. "What could go wrong with this approach?" "What am I missing?" "Play devil's advocate." It won't catch everything, but it catches the easy stuff you're too close to see.
Prioritization and frameworks. Feed AI your options and constraints. Have it help structure the tradeoffs. It's good at organizing messy inputs into something you can actually decide on.
Communication and alignment. Drafting stakeholder updates, exec summaries, launch comms. AI handles the structure; you add the nuance and judgment.
Competitive research. Quick landscape scans, feature comparisons, market positioning. Not perfect, but fast enough to inform early thinking.
The pattern: AI handles the drafting, structuring, and first-pass thinking. You handle the decisions, context, and taste.
The real shift
AI doesn't just make work faster. It shifts value from production to judgment. From creation to commitment. From activity to decisions.
The floor rises for everyone. What separates good product work is still the same: knowing when AI helps, knowing when it shouldn't be used, knowing what good looks like in your domain.
AI can't fake taste. And AI won't know the ins and outs of your customer like you. That's why PMs matter more, not less.
I'm still figuring this out
I don't have all the answers. I'm experimenting, building, learning what works. Some of this will probably be wrong in a year.
But here's what I know: the teams that adapt to this will move faster than the ones clinging to old workflows. And the PMs who learn to work with AI as a real partner — not a feature, not a tool, a partner — will have a massive advantage.
The future of product work is going to be fast and different. I'm here for it.
Want to go deeper? I'm building out a whole section on AI for Product Management with frameworks, tips, and lessons from actually doing this. Check it out.