Signals

A signal is an organization-scoped row created from a source publication (for example a municipal omgevingsvergunning). It is not the raw XML/JSON payload: that lives on raw_publications with an external_id and content hash for dedupe.

Fields you will seeCopy Anchor Link

• Title, summary, source URL, published / effective dates
• Status — how far the pipeline has got
• Signal type — from classification
• Current score — latest total, 0–100
• Project — set after entity resolution when a match is accepted

Child rows (not duplicated on the list) include classification output, locations, mentioned companies/people, identifiers (KVK, and so on), and score components.

What happens after ingestCopy Anchor Link

1. Classify — AI structured output: commercial relevance, category, stage, size, reasons. Source text is wrapped as untrusted data.
2. Qualify / score — scoring rules plus AI assessment. Below the ignore threshold, status becomes ignored.
3. Entity resolution — match or create a project and companies. Low-confidence matches are suggested, never auto-merged.
4. Enrich — contacts via Apollo when configured, with a generic info@ / main-phone fallback.
5. Ready — if the score meets the opportunity threshold, an opportunity is created. CRM sync is a separate step (Attio).

You can reprocess from a chosen stage (reclassify, re-score, re-enrich) without re-ingesting the publication.