Glossary

What is an evidence graph?

An evidence graph is a data structure that links each decision to the specific claims, evidence, sources, and confidence levels that produced it. In lending, it lets a team trace any AI-assisted decision back to exactly what was relied on — the foundation for explaining a decision to a regulator or auditor.

Why decisions need provenance

Regulators increasingly expect lenders to explain automated decisions with specific, accurate reasons. A logged output ('denied') is not enough; you need the chain from claim to evidence to source. The evidence graph is that chain, stored as a first-class record.

How Hadrian uses it

In Hadrian, each case decision carries its evidence graph, so the specific principal reasons for an outcome — and the documents and signals behind them — are reconstructable on demand. This is what turns an AI-assisted decision into a defensible one.

FAQ

Evidence Graph — common questions

How is an evidence graph different from an audit log?

An audit log records that something happened; an evidence graph records why — linking the decision to the claims, evidence, sources, and confidence that drove it.

Tamper-evident audit log Adverse Action Notice

The institution around the intelligence

See Hadrian run your case lifecycle — intake to close, every decision audited.

Governance-native case processing for lenders and regulated teams.

Book a live demo