Make Disappointment Your Instrument
Vohra's core move is a single survey question: how disappointed would you be if this product went away? If forty percent or more say "very disappointed," you have signal worth building on. For AI-native founders, this matters immediately. Agents can generate output at scale, but output is not value. The question to ask your early users is not whether the product is impressive — it is whether losing it would hurt. Impressiveness is a demo quality. Hurt is a retention quality.
Segment Toward the Believers
Vohra instructs founders to isolate the users who already answer "very disappointed" and study them: what they do, how they describe the benefit, what they had before. That segment is your real market. Distribution in a world of infinite software is about reaching exactly these people first and fast. The founder who has lived the customer's problem already has a hypothesis about who this segment is. The survey confirms or corrects it. Either outcome is useful; neither outcome is available without doing the work.
Split the Roadmap Deliberately
The method produces two parallel workstreams: deepen what the lovers already love, and remove the specific objections that keep persuadable users from crossing over. This is not compromise — it is prioritization with evidence. For founders shipping with agents this year, this split maps cleanly onto product architecture. Judgment calls about which objections are worth solving, and which signal a user who will never convert, are exactly the decisions humans must own. Agents can surface patterns in feedback at volume; only a founder can decide which pattern represents a real bet.
