Evidence that stays audit-ready by itself

The failure mode of manual evidence collection is a shared drive full of screenshots that nobody can map back to a control, half of which are stale by audit time. Fintra tracks two things that make evidence usable: coverage (does each control have the evidence it requires) and freshness (is that evidence still within its validity window). A control with expired evidence is treated as a gap, not a pass.

Good evidence tells a story. Fintra records lineage, the ordered sequence of events that back a piece of evidence, so instead of a single screenshot, an auditor sees the chain: the review that happened, the approval that followed, and the record that resulted. For governed AI actions, the lineage is the hash-chained ledger entry showing the decision and its context.

• Each evidence item links to the ordered events that produced it.
• Governed-action evidence is hash-chained, so a tampered link is detectable.
• Lineage turns "here is a file" into "here is how this came to be true."

The differentiator versus config-scanning tools: when an AI agent or automation acts, the Policy Decision Point returns a verdict and writes it to the evidence ledger. That decision is first-class evidence, proof a control operated on the automation layer, not a log line you reconstruct later.