Anthropic has announced Enterprise Frontier Safeguards, a new architecture meant to give companies stronger AI misuse monitoring while keeping activity data inside cloud infrastructure they control.
The offering, announced September 1, combines automated safety monitoring with customer-managed storage, encryption keys, access policies and audit logs. Anthropic says it will roll out the system in phases, beginning later this fall.
The company developed the safeguards with more than 100 organizations across sectors including finance, healthcare, manufacturing, law, retail and the public sector. It says the system was shaped around a recurring enterprise dilemma: regulated organizations need safeguards able to identify risky behavior that can unfold over multiple sessions, but many cannot accept a model provider storing sensitive activity data.
Under the new design, activity data used for monitoring can remain in a customer’s Amazon S3, Azure Blob Storage or Google Cloud Storage environment. Anthropic’s automated systems analyze a rolling window for serious misuse signals, while alerts are sent to the customer for review. The company says no Anthropic employee needs to conduct that review.
Our logs stay in a Wells-managed environment under Wells-managed keys.
— Munish Kumar Sharma, Wells Fargo chief information security officer
Enterprise Frontier Safeguards will be supported on Claude Code, Claude Enterprise, the Claude Platform, Amazon Bedrock, Google’s Agent Platform and Microsoft Foundry, according to Anthropic. The company says it will not charge for the controls, although customers hosting data in their own cloud accounts would pay their cloud provider’s usual storage and transfer costs.
The development is less a model launch than a deployment proposition: Anthropic is attempting to make advanced AI use more feasible for organizations whose security, privacy and compliance requirements have limited how they can adopt it.
