Glossary
Claude Enterprise tenancy
The isolated environment where your data stays yours and your AI agents are built and run.
A Claude Enterprise tenancy is the isolated Anthropic environment provisioned to a single organization, where the client data stays with the client and is not used to train models. It is BAA-eligible for HIPAA, and it is where Lama builds and runs your AI agents.
The tenancy is the boundary that makes the rest of the work defensible. Your documents, prompts, and agent logs live in an environment provisioned to your organization alone, not a shared pool, and they are not used to train Anthropic models. That isolation is the precondition for audit-ready AI agents, because access scoping and logging only mean something when the data sits somewhere you control.
For a regulated team the practical question is what touches patient or clinical data. Claude Enterprise is BAA-eligible, so a tenancy handling PHI can sit under a signed HIPAA Business Associate Agreement before anything sensitive is processed. Lama builds inside your tenancy rather than ours, which keeps the data residency, access controls, and human sign-off described in our security model under your authority across pharma and CRO work.
Common questions
Does Anthropic train on data in a Claude Enterprise tenancy?
No. Under Claude Enterprise, the prompts and documents you send stay yours and are not used to train Anthropic models. The tenancy is the boundary that keeps client data with the client.
Can a Claude Enterprise tenancy be covered by a HIPAA BAA?
Yes. Claude Enterprise is BAA-eligible, so a tenancy handling protected health information can sit under a signed Business Associate Agreement with Anthropic before any PHI is processed.
15 min. 5-day written diagnosis. No deck.