Updated 17 October 2026. I call it No Lanes Product Management: follow the problem across the boundaries it actually crosses. The point is not to do everyone’s job. It is to stop an important question disappearing between job descriptions.
Follow the consequence, not the handoff
An agent feature can connect product intent, data quality, tool access, commercial promises and operational support. A decision in one area can create work in another. Product leadership needs enough context to make those dependencies visible and bring the right people into the decision.
Consider an illustrative assistant that begins by drafting a recommendation. The team later proposes allowing it to execute the recommendation. That change affects more than the interface. It introduces questions about authority, approval, evidence and recovery. Calling it “just one more tool” understates the product change.
Following the problem means asking those questions without pretending to be the sole expert who can answer them. Engineering designs and tests the implementation. Legal assesses applicable obligations. Operations explains what happens after failure. Product helps keep the intended outcome and trade-offs coherent.
Clear ownership matters more, not less, when boundaries are crossed. No Lanes should not become No Responsibilities. Record who decides, who supplies evidence and who handles the next step. Collaboration is not an excuse for making accountability communal and therefore invisible.
| Question | Evidence or action |
|---|---|
| Intent | Outcome and limits of the feature |
| Consequence | Permissions, evidence and recovery |
| Ownership | Named decisions across disciplines |
The practical next step
Stay connected after release. Use failures and exceptions to improve the product definition, not merely to create tickets for another team. The customer experiences one product, however many departments built it.
Keep the resulting decision, its owner and unresolved questions in a short decision record. For claims about agent behaviour, connect the decision to evaluation evidence rather than relying on a successful demonstration alone.
Further primary reading: NIST AI Risk Management Framework. The review map and product recommendations above are my analysis, not a certification or a legal assessment of a particular deployment.
