Notes & sources.

Product

Scoring, rationales and the reconciliation figures

Referenced from the product tour.

1. Fit scores and rationales are AI-generated from public signals, not verified by a human before you see them. Treat each one as a well-informed starting point for your own judgment call, not a guarantee — review before you rely on a score or send an AI-drafted opener to anyone.
2. The reconciliation figures shown on that page are illustrative of a single metro search — the ratio of raw source hits to reconciled businesses varies by territory and by how many registries cover it.
3. Every account is scored; a full call-opener rationale is written for as many as the search’s time budget allows — the top slice, prioritised first. On rare occasions where a search runs long, a handful of accounts lower in the list may show a plain score and category summary instead of a full rationale. The ranking itself is unaffected, and it is visible in the results which accounts got the full write-up.

Data & sources

Where the layers come from, and what they promise

Referenced from the data page.

1. Row counts describe the reconciled search index and the datasets behind it; they change as sources are refreshed. Which layers run on a given search depends on the industry selected in Settings.
2. We do not guarantee that any particular data source will remain available or continue to be maintained. Public registries are published at their owners’ discretion: a provider may change its format, restrict access, stop publishing, or withdraw a dataset entirely, and we may add, change or remove sources in response. The set of layers behind any given search may therefore differ over time. Every result shows the sources that confirmed it at the time it was returned. This is stated formally in the Terms, section 6.
3. Curated award and recognition data is hand-verified in-house, per venue — not a bulk scrape of an awarding body’s site.
4. The permit record shown under Errata is a real source conflict in public federal data, reproduced to illustrate the reconciliation rule. State is assigned from the premises ZIP, consistently, and the record used is shown alongside the result.

Industries

Modes, labels and layer counts

Referenced from the industries page.

1. Mode names change per industry; the underlying two modes do not. Distributor search exists only where a licensed wholesale tier legally does.
2. Layer counts are the registered third-tier sources per industry. Federal layers cover every state even where a state’s own licence data is legally excluded or unpublished.

Pricing

How the credit arithmetic works

Referenced from the pricing page.

1. Full Territory Briefing scores and briefs every account in range, up to 240 per search, and multiplies that search’s credit cost by three. The multiplier is shown before the search runs.
2. Search counts shown as “roughly” assume standard searches at 5 credits. A deep search costs 15, and a Full Territory Briefing triples whichever depth you chose.
3. Fit scores and rationales are AI-generated from public signals and are not verified by a human before you see them.

Terms

The binding version