Describe a role and the engine returns accounts that lack it, ranked by revenue. Hand off any row to ZoomInfo with the target role pre-filled — or run an inline quick search. The Ideal Team Coverage chips are derived from your won deals (stored in indexes/contact_ideal_roles.parquet).

Describe any outreach list in plain English. The engine interprets event start-date windows, opportunity stages, scored prospects (event_prospect_matches and product_synthesis), and flags accounts whose best contact spans multiple product families. Use “no other open pipeline” to keep prospect-only accounts that lack additional CRM opps.

For each product line, which buyer archetypes were present on the deals we won versus the ones we lost. Only closed, pre-decision (OpportunityContactRole) records count — contract and delivery roles exist only on deals that already closed, so including them would score the outcome against itself.

Measure one account against a product line's blueprint. A filled seat names the contacts we already have (ranked by deal influence, then recency); a vacant seat hands back the titles to search for.

One row per human across 15 behaviour domains — events, webinars, badge scans, papers, newsletter, marketing email, website, subscription usage, surveys, polls, replies, forms, speaking, deals and qualified interest — rolled into four dimensions and stacked month by month back to 2010. Scores are read against what we could actually observe about each person, so somebody we hold no channel for reads as unobserved rather than disengaged.