Trust
Data provenance
Every research finding, candidate, and score in Leead.io can be linked back to its sources. Provenance is not a promise — it is a structured record of where information came from, when it was observed, and how it was extracted.
What we record
- Source URL and domain — the original and canonical URLs where information was found.
- Observed content — an excerpt or structured evidence extracted from the source.
- Retrieval time — when the source was observed, so you can judge freshness.
- Extraction method — how the evidence was derived, such as public-web research or API response.
- Independent source count — how many distinct domains or publishers support a finding.
- Verification state — unverified, partially verified, disputed, or rejected, with a human-verified state available for authorized reviewers.
Why it matters
Sales and marketing teams make better decisions when they can see the evidence behind a record. Provenance helps you understand whether a contact title was observed on the company website, a company size was reported in a registry, or a fit score was derived from multiple consistent sources.
Verification is not AI confidence
AI can surface evidence and propose verification states based on independent sources, but only an authorized human can mark a record as verified or rejected. Confidence scores are model estimates; verification is a deliberate review action.
Evidence classification
Repository Confirmed — provenance models, evidence records, source deduplication, content hashing, independent source counting, and verification state machines are implemented in V1.1.