Prototype Before the Sprint
Prototype Before the Sprint
Part of my routine as a product manager is building quick prototypes of feature ideas with AI tools like Lovable and Claude before they go into sprint planning. It has changed how I write PRDs and how conversations with engineering and design go.
A prototype answers questions a document cannot
A PRD describes what a feature should do. A prototype lets people click through it. Questions that would take a long thread to settle in a document, like what happens on this screen when there is no data, become obvious the moment someone tries it.
Cheaper ideas mean more ideas get tested
When a rough version takes an afternoon instead of a sprint, I can try more than one approach to the same problem. That makes it easier to drop weak ideas early, before anyone has committed engineering time to them.
It is a conversation starter, not a spec
A prototype built with AI tools is not production code, and I do not present it as one. Its job is to give engineering, design, QA and go-to-market teams something concrete to react to. The real design and build still happen with the team, with their judgement on how it should work.
Show it to customers early
A clickable prototype is also a good research tool. Customers find it easier to react to something they can see than to a description, and their reactions often point to edge cases the team had not considered.
What the Curious PM cohort added
Building AI products in the Curious PM cohort pushed this further. It is one thing to mock up a screen. It is another to get a working AI feature in front of people and see how it behaves with real inputs. That has made me more confident bringing working prototypes, not just wireframes, into planning.
Prabhakar skipped presentations and built real AI products.
Prabhakar Arya A was part of the August 2026 cohort at Curious PM, alongside 24 other talented participants.