AI disclosure
Last updated 27 October 2026
What Hub AI does
Hub AI uses large language models to read new email and produce suggestions: a priority score with reasons, summaries, reply drafts, task suggestions, follow-up reminders, a morning brief, answers to questions about your own mail, and phishing and prompt-injection warnings.
What it never does
Hub AI never sends, deletes, moves, archives or unsubscribes from email, never answers meeting invitations, and never makes payments. A person always takes the action.
AI can be wrong
Suggestions and warnings can be incomplete or wrong. Read drafts before you send them, and don't rely on a missing warning as proof that an email is safe.
Learning
Hub AI can propose "memories", such as your usual greeting or which senders matter to you. A proposed memory has no effect until you approve it, and you can edit or delete it at any time.
Our own assistants
On this website and on our support channels you may talk to an AI assistant. It always says it is an AI, it can hand you to a person at any time, and it never takes card details.