What problem does it solve? Niche B2B lists come out tiny because database keyword searches miss good-fit companies that aren't tagged with the obvious keyword. This Skill works bottom-up from ~10 known-good seed companies: it fingerprints how they actually appear in the company database, expands them through lookalike engines, mines high-precision filters, and pulls and AI-qualifies the full market at scale. ## Core Features & Use Cases - Seed fingerprinting and lookalike expansion: Discovers how seeds are tagged in Prospeo, then generates candidates via Prospeo company_lookalike, Exa findSimilar, and Parallel.ai entity search. - Filter mining with a precision scorecard: Mines industries and keyword n-grams from confirmed fits, then scores each candidate filter on volume and sampled precision before the wide pull. - Scaled qualification and verification: Scores thousands of companies with a cheap OpenAI model, verifies live websites to remove dead or parked domains, and produces a client-facing HTML transparency report with a verified-email TAM ceiling. - Use Case: A client says their medical-groups list of 200 companies should be much bigger. Feed 10 known-fit domains in, and the pipeline returns thousands of AI-qualified, website-verified companies plus a report showing exactly how the list was built. ## Quick Start Ask the AI to expand a list by running the list-expander pipeline starting with fingerprint.ts on your seed domains, for example: expand these 10 seed companies into a full qualified TAM for US multi-site medical groups.
