Goldie

Founders and product teams burn weeks doing market research by hand — competitor scans, TAM math, pain-point interviews — and then still ship the wrong thing. Goldie’s thesis: an AI agent can run that whole loop in hours, kill weak ideas before you fall in love with them, and return a ranked opportunity map with the supporting evidence attached.

Built in two weeks end-to-end as both a real product concept and a proof of the AI-native venture loop it sells. A Python wrapper around OpenAI’s o3-deep-research with web search produced a 131KB ethnography of the UX-research market — TAM/SAM/SOM, a 15-competitor teardown, an AI-maturity matrix, and primary-source citations — which became both the demo content and the validation for Goldie itself. On top of that, a Lovable-built landing page and a 13-screen interactive demo walk a visitor through the agent’s behavior: pose a problem, watch deep research run, see opportunities scored against nine strategic frameworks (Blue Ocean, YC filters, Zero to One, Porter’s Five Forces, Disruption Theory, 7 Powers, Working Backwards, Hype Cycle, Unfair Advantage), and land in a results dashboard. A Clueso-produced hero video sells the whole loop in 60 seconds.