Developing Enterprise Frontier Safeguards with our customers

September 1, 2026 | Source: Anthropic News (community RSS)

Tags: Enterprise Frontier Safeguards, Anthropic, enterprise AI, data privacy, Claude Enterprise, Amazon Bedrock, cloud security

Anthropic's Enterprise Frontier Safeguards (EFS) resolves the data privacy vs. safety monitoring conflict blocking regulated-industry Claude adoption by storing interaction data in customer-controlled AWS, Google Cloud, or Azure infrastructure — built with 100+ enterprise customers.

Details

Enterprise Frontier Safeguards (EFS) directly addresses what blocked Fable 5 adoption in regulated industries: the 30-day data retention policy needed for cross-session misuse detection conflicted with zero data retention requirements in financial services, healthcare, and law. EFS stores all Claude interaction data in cloud infrastructure fully controlled by the customer — not Anthropic — while still enabling the safety monitoring required to catch coordinated fraud, credential theft, and multi-session cyberattack patterns that span accounts. Anthropic argues single-interaction analysis is insufficient for sophisticated threats; cross-session correlation requires storing data for a meaningful window, but it doesn't have to be stored at Anthropic. EFS was developed with more than 100 enterprise customers in financial services, healthcare, manufacturing, telecom, law, retail, and the public sector, alongside cloud partners AWS, Google Cloud, and Azure. It rolls out in phases from fall 2026, covering Claude Code, Claude Enterprise, Amazon Bedrock, Google Agent Platform, and Microsoft Foundry. Anthropic explicitly states it has never trained on enterprise data without permission. The effort involved CISO-level collaboration through the Analysis and Resilience Center for Systemic Risk (ARC).