What problem does it solve? Sales and marketing teams often need to know whether each company on a prospect list mentions a specific keyword (a certification, technology, or service) on its own website, but no structured data provider sells that field. This playbook answers that question at list scale using one Google site: query per keyword through a SERP API, then turns verified hits into copy-ready personalization lines with evidence URLs. ## Core Features & Use Cases - Keyword presence detection: Runs a single literal-phrase site: query per domain, filters results deterministically, and applies a model judge to decide whether the company genuinely talks about itself. - Gated personalization output: Produces site_keyword_hit, a confidence-gated site_keyword_line, an evidence URL, and a pre-rendered sentence that renders cleanly even on abstain rows. - Cost-controlled pipeline: Blended cost of roughly $0.59 per 1,000 rows, with a literal-phrase credit gate that prevents wasted model calls. - Use Case: Given a list of 500 B2B domains, filter which ones mention SOC 2 on their own site, then upload only the high-confidence rows to a sequencer with a Noticed-sentence personalization variable. ## Quick Start Ask the assistant to check which domains on your list mention a specific keyword on their own website using a Google site: search and return a personalization line for each verified hit.
