PMing in the AI Era: From Chaos to Specs AI Can Actually Build 🤖🔥

We’re shipping faster than ever… and also explaining way more why things didn’t turn out quite right 😄

This Product Tank brings two sides of the same story:

👉 How to survive (and function) as a PM in AI chaos
👉 How to write specs that AI can actually build from

Talks:

🎤 Jindra Regál ()
You Are the Ultimate Product: Creating Value in Non-Tech Fields

In the AI era, everything gets built multiple times a day. So what problems are still worth solving?
After working with companies like Avast, Productboard, Better Stack, Ataccama, and BottleCap AI, I found the most meaningful work came from solving small, practical problems for real people and businesses.
A talk about applying product thinking outside IT companies, expanding your perspective beyond software, and learning to treat yourself, your skills, curiosity, and adaptability, as the ultimate product to manage.

⸻

🎤 Jakub Elias (Similarweb)

PMing in the AI era
PMs are having an identity crisis. Designers prototype with AI. Engineers spec with AI. You hear about CEOs shipping features from the back of an Uber. So what's left for you? Turns out, still quite a lot.

This talk is about what the job actually becomes when building gets cheap and deciding gets hard. How the trio holds together when everyone's on everyone else's turf. How to keep a roadmap alive when the ground shifts every few weeks. And lastly, what's now expected of a PM - being a real builder.

⸻

🎤 Dominik Ilichman (Productboard)

AI coding agents have compressed software delivery like never before — features that used to take weeks now ship in days. But here's the catch: agents don't improvise. They don't fill gaps, ask clarifying questions, or read between the lines. Ambiguity doesn't get resolved — it gets built wrong. The bottleneck has moved from engineering capacity to clarity, and the product specification is now the most critical artifact a PM produces.

In this talk, Dominik explores how the PM's role is fundamentally shifting as a result. We'll look at what great AI-ready specs look like in practice: grounding them in your codebase and real customer feedback, using prototypes as a diagnostic tool, and why the spec — not the code — is becoming your team's true intellectual property. He'll also show how tools like Productboard Spark are flipping the traditional workflow — instead of the PM writing and AI helping, AI drafts and the PM steers — making it possible to go from opportunity to delivery-ready specification in a fraction of the time.

If you're a PM figuring out how to stay high-leverage in the age of AI, this talk is for you.