The definition did not change, which is the point Cagan is making
In August 2026 Marty Cagan published a piece titled "A Fresh Definition of The Product Role." Read closely, it does not contain a new definition. It takes three skills Cagan credits to Benedict Evans, from an article of Evans's called "Most People Aren't Tool Builders," and maps them onto the framework SVPG has used for two decades.
Recognizing the general problem behind a pile of specific pain becomes problem discovery. Being able to create a good solution, as distinct from being good at using someone else's tool, becomes value risk. Understanding what a solution means across sales, marketing, finance, compliance, and legacy systems becomes viability risk. Cagan's conclusion is that AI changes the thresholds and not the problem.
He is right that the skills survive. The part I would argue with is the weighting, because a definition that treats the three as one job in roughly equal parts hides the only thing that actually moved.
The three have completely different exposure
Solution discovery is the most exposed thing in product work. Not because a model has taste, it does not, but because the expensive half of that skill was producing credible candidates at all, and that half is now close to free. What remains scarce is choosing, and choosing is a different skill with a different training path. The profession has twenty years of people who got good at generating options and almost nobody who got deliberately good at killing them.
Problem discovery is intact and its input changed underneath it. The skill was always bottlenecked on how many pains one human could personally hear, which is why the canonical advice is a weekly interview cadence: it is the most listening a person can sustain. A listening agent hears every ticket, call, churn survey, and rage click. Same skill, different sample size, and a skill practiced on twelve conversations a quarter behaves differently when it is practiced on everything.
Business viability barely moved, and that is exactly why it is becoming the majority of the job. When the other two thirds get dramatically faster and this one does not, it takes over the calendar by arithmetic alone. Nobody is hiring for that yet. Product job descriptions are still weighted toward the third that is deflating fastest.
What I do with this on my own product
On Heidi, the agentic platform I build, this is the reason the hard engineering is in the trust layer rather than the builder surface. The builder surface is solution discovery made cheap, and anyone can ship a version of it in six months. The parts that do not compress are the ones that map to viability: permissions, reversibility, guard rails, and receipts for what an agent did and why, so that a company can say yes to a long-running agent without a person in the room for every act. That work is unglamorous, it is most of the roadmap, and it exists because consent inside an organization is the constraint that AI did not touch.
The same logic applies to a career. If you are betting on tool fluency, you are betting on the deflating column. Useful, necessary, and aimed at the third that is getting cheaper every quarter.
Sources: A Fresh Definition of The Product Role, Marty Cagan, SVPG, and the podcast episode with Benedict Evans that Cagan links for the paywalled original.