Worked example: sales & GTM targeting

Domain matching gets a form-to-account rate around 63%. Resolving the company field instead gets 84%, and the difference is almost entirely companies whose domain is not their name.

Size the market from the register

A vendor estimate is someone else's segmentation. GET /v1/search counts the actual population: legal form, sector, geography and vintage, from the register.

$ curl -s "https://api.spotit.ai/v1/search?country=DE&nace=49.41&headcount_min=50&status=active&count_only=true" \
    -H "Authorization: Bearer $SPOTIT_KEY"

{"count": 3184, "as_of": "2026-09-05"}

Resolve at capture

Put the resolve call in the form handler, not in a nightly job. Routing then sees a legal entity rather than a free-text string, and the rep gets an account rather than a lead to disambiguate.

1. Resolve company plus the email domain as a hint. p50 is 38 ms — inside the budget of a form submit.
2. On auto_accept, write entity_id to the lead and look it up against your account table.
3. On review, keep the lead but flag it. Do not block the form on ambiguity.
4. On no_match, accept the lead as typed. A missing match is not a reason to lose a hand-raiser.

Suppress by ID, not by domain

Suppression on domain misses every subsidiary that shares a parent's website and every account that changed domain after a rebrand. GET /v1/entities/{id}/relations gives the group, so a suppression list built on IDs covers the whole structure.

select l.lead_id
from leads l
where l.spotit_entity_id not in (
  select r.related_entity_id
  from spotit.relations r
  join crm.accounts a on a.spotit_entity_id = r.entity_id
  where r.kind in ('parent','subsidiary','branch')
)

What it moved

Related: CRM hygiene and dedup, and the matching guide for choosing the auto-accept band.