Your data
We build on the business and API tiers of Claude and other model providers, which by default don’t use your data to train their models. Integrations run in accounts you own. During a pilot we can run them on our own Claude backend, then move them to your accounts at handover.
Access and governance
We ask for the minimum access each integration needs, set who approves sensitive actions, and remove our access when the engagement ends. We work under your NDA on request.
For company-wide rollouts, governance is part of the project: admin settings, user permissions and the written policy on which data staff may use with AI. More on this: how employee AI use feeds training data.


