Qualify a prospect. Get one verdict and a drafted response. Sonar decides. It doesn't hand the question back.
Studios hit your inbox every week. Most aren't right. Screening them by hand takes time you don't have.
Wrong genre. No funding. Wrong moment. A generic AI writes a confident paragraph about all of them anyway.
Output record pastes straight into col-15 of the MASTER sheet — action + warm intro first, score breakdown supporting.
PURSUE · primary-vendor Action: Email Dana Reyes (Exec Producer) — October ship date, no audio team yet. The in: Warm — Dana Reyes, met at GDC 2025 (CRM). Score: 69/107 · Tier 2 — audio gap 22, active dev 12, horror/narrative 10; PURSUE via live catalyst. Catalyst: project — announced October ship window.
Six worked decisions in examples.md — one per verdict, plus a not-found failure. Fire the prompt, get the same output.
Built by the person running the workflow.
CCI does BD the same way Sonar does — relationship-driven, ICP-disciplined, no cold-spray. This is the operator built for the workflow I'm already running. Then Shipper — the launch operator built the same day — dogfooded it and produced this deck.
Swap the rubric, catalog, and ICP for your own. The verdict logic carries over unchanged.