Why isn't the roadmap the CPO's job anymore?

Because it rationed engineering, and engineering is no longer the constraint. SC Moatti's survey of more than 1,500 CPOs found product cycles shortened from months to days and the bottleneck moving to go-to-market and adoption. A plan for spending a resource that isn't scarce doesn't steer anything.

What went wrong at Atlassian?

Per Jason Lemkin's SaaStr write-up of Sherif Mansour's talk, about ten projects staffed with hand-picked AI builders moved fast for a few weeks and then slowed, because nobody was making the calls on what to build. PMs who had started coding drifted back to customer context and decisions, which unblocked the teams.

What are the five parts of the CPO's operating model?

The substrate engineering builds so anyone can ship safely, the listening loop that gets customer evidence to builders, the eval bar that defines done, the org shape with the Product Builder as the unit, and the margin, since every AI feature has a variable cost of goods.

Should a company merge the CPO and CTO into a CPTO?

Either shape works if one person owns the path from customer outcome to shipped software. In October 2026 Dayforce merged the seat under Jake Varghese, while Guidewire hired Dan Brown as CPO and kept Diego Devalle over product development. The title matters less than whether a CEO can name the owner of the loop.

What should a CEO ask a CPO instead of the roadmap?

Show me the loop. For the last five things shipped: who decided they should exist, what customer evidence they had, what bar they cleared, and what they cost to run. A CPO who can answer is running an operating model.

What Does a CPO Own When Prototypes Are Cheap?

What does a CPO own when prototypes are cheap?

THE SHORT ANSWER

The operating model, not the roadmap. The roadmap was a rationing device for scarce engineering. When a working prototype costs a day, rationing stops being the hard part, and what's left is deciding what should exist, proving it worked, and making it pay for itself. Atlassian showed the failure mode in October 2026: per Jason Lemkin's SaaStr write-up, Sherif Mansour's team staffed about ten projects with hand-picked AI builders, moved fast for a few weeks, then stalled because everyone was rowing and no one was steering. The CPO owns five things instead: the substrate engineering builds so anyone can ship safely, the listening loop that gets customer evidence to builders daily, the eval bar that defines done, the org shape with the Product Builder as its unit, and the gross margin of every AI feature. Whether the seat is a CPO or a merged CPTO matters less than whether one person plainly owns that loop.

The roadmap rationed engineering. When a working prototype costs a day, that's no longer the hard part, and Atlassian's October 2026 experience shows what happens when speed arrives without steering: about ten AI-builder projects flew for a few weeks and then stalled, because, in Sherif Mansour's words as Jason Lemkin reported them, everyone was rowing and no one was steering.

So the CPO owns the operating model. The substrate, the listening loop, the eval bar, the org shape, and the margin, each of which gets its own part in the series. Craig Unsworth's warning applies to the ratio debate that usually follows: it's possible to make everybody faster and the company no faster at all. The CPO's question is where decisions get made and how fast the evidence reaches them.

The full argument is Somebody Has to Steer: The CPO Now Owns the Operating Model, part 1 of The CPO Transition. The handbook chapters underneath it are Kill the Roadmap, Engineering Builds the Substrate, Not Features, and Gross Margin Is Your Job Now.

Last reviewed 2026-10-14 · 1 min read

What does a CPO own when prototypes are cheap?

SOURCES

THE LONG VERSION

RELATED ANSWERS