Research and Enrichment
The two agents that turn an email address into a record with enough context to judge, and how they decide what a fit looks like.
Enrichment fills in what is already known: given an email and a domain, it fills the known fields and merges the form, the webhook and the CSV row into one lead instead of three half-finished copies. Covered in Enrichment.
Research is the judgement: it reads the account's public pages against your ICP to decide what this company is and whether it matters.
What Research Adds
| Output | Example |
|---|---|
| Segment | "B2B SaaS · Compliance", more specific than an industry code |
| Context | What the company does, in one line a rep can read before a call |
| Account news | Funding, hiring and public launches, each attached with the source it came from |
| ICP verdict | A qualification decision against your definition, with the evidence behind it |
Where It Looks
The enriched record and the company's public pages, judged against the ICP you defined in the Leads Hub.
Teams also keep a prose ICP in the Vault at
product/icp.md, for the nuance a set of weights cannot carry: "we win in
compliance-heavy SaaS, we lose in agencies". See
what each agent reads.
Failure Modes Worth Knowing
Free-provider emails. No company domain resolves, so research has little to judge and fit has little to score.
Ambiguous domains. Holding companies, agencies and multi-brand groups are
the usual source of a company match that looks right and is not. If a subsidiary
matters to you, say so in product/icp.md and spot-check those records.
Very new companies. Nothing public exists yet. A thin record here is accurate, not broken.