Ask a founder where their idea came from and the honest answer is usually "it annoyed me personally." That's a fine source, but it's one person's annoyance. The internet is currently full of people describing their annoyances in public, in detail, for free — in one-star reviews, forum threads, rising search queries, and questions nobody's answering well. Reading all of it is a research job no human wants. So we gave it to agents.
Five readers, one brief
The Discovery team runs five research agents in parallel, each mining a different vein: rising search trends, social discussions and complaints, Q&A sites, competitor gaps, and vertical data sources. Separately, each produces leads. Together, they converge on the leads that show up from multiple directions at once — which is the signal worth trusting. A trend that's also a complaint that's also an unanswered question is not a coincidence.
The output is a written Opportunity Brief, and its format is deliberately strict:
- The problem, stated as the people who have it state it.
- The audience — who exactly, and where they gather.
- The evidence — links, quotes, and numbers the agents actually found. No evidence, no brief.
- An MVP scope small enough to build and ship this week, not this quarter.
- A confidence rating the agents assign and justify.
The button that matters
Every brief carries a Build-this button. Pressing it hands the brief straight to the Builder team as a fully-formed spec — the research becomes the plan without a copy-paste step. Several of the products in our own Showcase started life as briefs nobody on the team thought of first.
Discovery runs once, on demand, or continuously on a schedule that keeps refilling your brief queue. And it's entirely optional: if you already know what you want, the text box is right there.
← All posts