Skip to content
Pondara Pondara docs

Export the audit trail

The audit trail is the regulator-ready log of every state transition that touches a learner record, a scenario, or a validation. It exists so that, when a regulator or an internal compliance review asks how a particular scenario reached Live — or how a particular learner’s competency record was built — you can answer with a timestamped record, not a reconstruction.

What the trail captures

  • Scenario lifecycle transitions. Draft → Pending validation → Validated → Live → Retired, in both directions. See the scenario lifecycle.
  • Cosmetic edits. Typo fixes and copy polish on Live scenarios, with the editor’s identity.
  • Substantive forks. When a Live scenario is forked into a new draft, both the fork event and the new draft are logged.
  • Trainer verdicts. Every Accept / Adjust / Reject on AI-extracted evidence, with the trainer’s identity and the evidence row it applied to. See evidence and verdicts.
  • Member and role changes. Invites accepted, role edits, and revocations.

Each entry carries a timestamp, an actor (the user who acted), an action (the verb), and a target (the scenario, evidence row, validation draft, or member affected).

Step 1 — Open the audit page

From your home, go to the audit page. You’ll see a chronological list of recent entries, newest first, with filters for date range, actor, action type, and field.

Step 2 — Pick a date range and format

Choose the date range you want to export. Pick a format: CSV for spreadsheets and most regulator-facing tooling, JSON for downstream ingestion. Click Export.

Step 3 — What the export includes

The export contains every audit entry in your org for the period, across all fields and cohorts unless you’ve narrowed the filters first. Each entry carries the timestamp, actor, action, and target, plus a stable id you can quote in correspondence.

The export does not include dialogue content (the words a learner and the AI sparring partner exchanged inside an engagement). Dialogue export is a separate workflow that goes through a privacy review on request — talk to your customer success contact.

When something goes wrong

  • The export is empty. Your date range may sit outside any org events. Widen the range, or remove the filters first.
  • The export is huge and slow. Narrow by field, by actor, or by action type before exporting. A year of audit entries across a large org can produce hundreds of thousands of rows.
  • An entry looks wrong. Audit entries are append-only; they are not edited or deleted. If the underlying action was wrong (a verdict in error, a wrongly-promoted scenario), the corrective action lands as a new entry — Reject, Retire, or fork — and both show in the trail.
  • You need dialogue content for a regulator. Contact your customer success contact. Dialogue export is gated by a privacy review and may need learner consent depending on jurisdiction.