Account Mapper: AI account enrichment that gives outreach actual ammo

Personalized outreach is dead the moment the prospect can tell it’s a template. Account Mapper exists so personalization has something true to say. Part of the complete AI GTM engine guide.

What Account Mapper does

Takes the account-fired event from Intent Watcher and runs a waterfall enrichment: firmographics, people (decision-makers + champions + influencers), verified emails, tech stack, recent funding, hiring posts, news mentions. Stops at the first source with high enough confidence — you don’t spend credits enriching things you already know.

Why waterfall logic matters

Single-source enrichment fails silently. ZoomInfo misses 30% of companies; Apollo misses different ones; Clearbit different again. A waterfall pattern — hit source A, fall through to B if confidence is low, then C — covers 85–95% of any ICP at a fraction of the cost of paying for all three.

Account Mapper runs the waterfall pattern Clay popularized but does it inside your AI GTM engine, with the providers and confidence thresholds you control.

ICP scoring

Every enriched account gets a score against the ICP you defined at sprint kickoff. Score uses signal type (Intent Watcher), firmographic fit (Account Mapper), and historical conversion patterns (Revenue Pulse data, once it’s running). Accounts below threshold get logged but not contacted. Accounts above threshold flow to Sequence Composer.

Stack & integration

Built on Clay waterfall (data layer), Anthropic (reasoning + scoring), Supabase (account state). Outputs to whatever CRM you use. Apollo, ZoomInfo, BuiltWith, Crunchbase, and your own first-party data are all configurable as providers.

Common failure modes

• Enriching the full TAM (burns credits, doesn’t produce better outreach).
• No confidence threshold (sends low-quality data to Sequence Composer, kills reply rates).
• Scoring on title alone (a VP at a 50-person company is not the same buyer as a VP at a 5,000-person company).

FAQ