What Anthropic receives: Only the data you explicitly submit as part of an AI task — for example, a document you've uploaded for summarization, or a client note you've asked Bo to draft a response to. Anthropic does not receive your full client database or any passive background sync data.
What Anthropic does not do: Anthropic's API terms prohibit training models on customer data. HNDL has signed Anthropic's Data Processing Addendum (DPA), which contractually confirms: no model training on your data, 48-hour breach notification, deletion of processed data within 30 days of termination.
US data transfer: AI processing constitutes a cross-border data transfer under PIPEDA. We disclose this transfer, obtain consent during account setup, and rely on the Anthropic DPA as the contractual safeguard. This is the same approach used by every Canadian business that uses any cloud AI tool, and it complies with PIPEDA's accountability principle.
You can disable AI features: If you prefer not to use AI processing for any client, you can skip AI-powered features. Non-AI features remain fully available.