Trust and privacy

What is a no-training clause?

A no-training clause is a contract term in which a provider promises not to use your data to train or improve its AI models. It matters for customer email because messages contain personal and business details that should be used to answer the customer, not to teach a model used by others.

The promise can sit in the terms of service, the data processing agreement or a product’s privacy settings. Consumer AI apps often train on conversations unless you opt out; business and API terms usually rule it out by default.

Why it matters

For an AI email tool there are two layers to check: the tool itself, and the AI providers it uses, its sub-processors. A clause from the tool is only as strong as the terms it holds its providers to. Look for the exact wording, “will not use customer content to train”, rather than a general reference to privacy.

Keeping what a tool learns about your business within your own account, to write your replies, is a different thing from training a shared model, and a good vendor explains which is which.

An example

A business asks a drafting tool whether its AI providers can train on the emails. The tool points to the providers’ business terms, which forbid it, and to its own DPA.

In ReplyPen: The AI providers ReplyPen works with operate under business terms that forbid training on your content. Details on where your data goes.

Related terms

ReplyPen drafts replies to customer email in Gmail and Outlook, with a Check note under each Draft showing what it relied on.

We're opening to new teams soon. We'll email you when your spot is ready.

Last updated September 2026