AI-tool monitoring notice

• Your full prompts. The complete text you send to covered AI tools.
• The full AI responses. Everything those tools return to you.
• Where AI content surfaces in your work. Metadata only: the page where it surfaced, the time, etc.

To keep sensitive company data from leaking into AI tools, and to trace where AI-generated content lands in real work. This supports cyber-insurance readiness and security and audit questionnaires. It is governance of the AI data boundary, not performance surveillance. There is no automated scoring or discipline; a human reviews anything that matters.

• Block your work. AI content that lands in your emails, docs, or tickets is never blocked. You may be asked to acknowledge it, but it always goes through.
• Score or discipline you automatically. Nothing here is an auto-score or an auto-action. A human reviews anything that matters.
• Read what you type outside AI tools. Outside covered AI tools, Zeflin checks one thing only: whether content surfaced onto a page matches AI content captured for your employer. It keeps a record only when it does.

Captured content is retained for 90 days, then hard-deleted. It is permanently removed, not archived.

The prompts and responses captured about you are your personal information. You may access, delete, or correct it through your employer's DSAR process, answered within 45 days.