AI-Proposed Remediation Actions

Finding a violation was never the slow part - acting on it was. Now each violating entity carries an action button that launches an AI agent, applies the fix, re-checks the rule, and reports the outcome back onto the same card.

What you can do

• See it before it runs - description, confidence, and priority on the card.
• Fix one entity or the whole rule - Apply rule-wide shows the exact scope first.
• Watch the agent work - every fix runs in its own chat, linked from the card.
• Decide what gets offered - a rule's Action Required field drives the fixes.
• Decline what you don't want - for that entity and no other.

Entities under one rule can get different fixes, up to five per rule, and a daily refresh picks up new violations. Fix actions and acknowledgements stay separate: acknowledging hides an alert, a fix action corrects the data.

Find them in the violations table under Data Governance.

Check the documentation →