Concentric AI adds Claude compliance auditing integration

Concentric AI has introduced an integration with Anthropic's Compliance API for Claude, allowing customers to audit activity on the Claude platform.

The integration works with its Semantic DLP product and records prompts, responses and attachments across both browser and app versions of Claude, including historical activity.

The move addresses a growing challenge for companies using generative AI tools in day-to-day work. Many security products were designed to detect malware, network intrusions and external attacks, but Concentric AI argues that generative AI also creates risk when employees with legitimate access to internal data use that information in unintended ways.

This can expose sensitive corporate material without clear oversight, especially when staff paste confidential information into AI prompts to speed up tasks. That, in turn, can create compliance issues, liability concerns and obstacles for internal investigations.

Concentric AI says its Semantic DLP service gives organisations visibility into the public generative AI tools used by staff, including unsanctioned applications often described as shadow AI. It also allows security teams to apply rules to block or redact sensitive information before it is shared in prompts.

Audit trail

Under the new integration, the system logs Claude platform activity for audit and review, including records of policy violations and the content exchanged through the service.

The product can also identify sensitive data generated in model responses, giving organisations an additional layer of oversight beyond monitoring what users submit. That information can then be used alongside Concentric AI's broader data security governance tools to identify insider risk and reduce leakage of sensitive information.

The launch reflects broader demand from companies that want to deploy generative AI tools but have not fully mapped where sensitive data sits across their systems or whether it is properly labelled. In those circumstances, employees may have access to more confidential information than they need for their roles, increasing the risk of exposure through AI services.

According to Concentric AI, Claude has gained traction among developers, professional users and large organisations. The company is positioning the integration as a way to give those businesses more visibility into what data is shared with the platform and how it is used.

Karthik Krishnan, founder and chief executive officer of Concentric AI, described the launch as part of a broader effort to support AI adoption while maintaining oversight of sensitive information.

"With this new integration, Concentric AI Semantic DLP empowers organizations to harness the revolutionary capabilities of the Claude Platform while meeting their need for robust data security governance. This release continues our commitment to providing organizations with a single end-to-end platform that protects all their sensitive data, wherever it lives and however it travels. Now, users can enjoy Claude's productivity and innovation while maintaining visibility and protection for their most valuable asset - their data. Supporting this use case shifts our AI-driven data security governance platform from risk reduction and compliance to a business enabler as well," said Karthik Krishnan, founder and CEO of Concentric AI.

Automated access to usage data is central to making oversight practical at scale. Without it, businesses would have to rely on periodic checks and manual review processes, which can be difficult to sustain as AI use expands across departments.

Anthropic's Compliance API gives third-party providers a way to capture and review Claude activity for governance purposes. By linking that access to Semantic DLP, Concentric AI aims to place Claude usage within the same monitoring and policy framework companies use for other forms of sensitive data handling.

The integration is now available.