AI Governance

Responses are generated by Anthropic's models. Anthropic is our AI provider today, and our privacy policy names it as a service provider.

When you run a tool or workflow, we send your request and the context for the active client, such as the voice samples, boilerplate and key messages you saved in the Client Profile. Some tools also search the web. When they do, your query terms go to that search function.

We don't use your account or client content to train AI models, ours or our provider's. Content sent for processing is used to generate your response. Each client's data is scoped to your own account or organisation and is not visible to other customers.

• The workflow steps are written to work from the facts you supply and to leave out facts you did not give them.
• Research steps that search the web are set to link the pages they used, so you can open each source.
• Writing is checked against our house style, which rules out a list of common AI phrasings, before you see it.

Every result is a draft. You read it, check it against your own sources and decide what goes to a journalist or a client. Writing Review gives you feedback on a pitch or press release before you send it.

If you want to ask about how we handle AI or data, write to hello@theprassistant.ai.