AI limitations and responsible use

Where AI appears

• Credential extraction from an uploaded document, when a credential is first created.
• AI Compliance Brief on the Dashboard.
• Explain Compliance on the Compliance Matrix and on individual action items.
• Renewal review summaries and drafted employee communications.
• Ask ComplyNestly, Copilot priorities, and Change Intelligence.
• Executive Compliance Brief and report narratives on the Reports page.

What AI can get wrong

• Extraction can misread a date, number, or authority name on a scanned or low-quality document.
• Summaries can emphasize the wrong thing or omit a detail that matters to you.
• Narratives describe what your data shows, which is only as accurate as the data itself.

Rules for relying on it

• Treat AI output as a draft to review, never as an authoritative record.
• Confirm anything you will send externally against the underlying data.
• Do not treat any AI output as legal, regulatory, or licensing advice.
• Edit drafted communications before sending them.

What AI never does

• It never approves or rejects a renewal or submission.
• It never edits or deletes a record on its own.
• It never sends a message without an explicit send action from you.