Reco launches Claude security integration for enterprises

Reco has launched a bidirectional security integration for Claude, aimed at enterprises using Claude across employee and developer environments.

The product is designed to give security teams oversight of activity in Claude Enterprise and Claude Platform, while also letting them use Claude to investigate risks identified through Reco's data graph.

As businesses adopt generative AI tools for day-to-day work, software development and agent-based workflows, security teams face a new set of governance challenges. These include limited visibility into who is using AI systems, what data they can access, which permissions have been granted, and how connected agents interact with other business applications.

The integration links Claude activity to the Reco Graph, which maps applications, identities, permissions, workflows and data paths across an organisation. By connecting those elements, security teams can identify overpermissioned agents, unmanaged access, stale API keys, risky connections and the potential impact of a compromised tool or account.

It is intended for organisations using Claude in two main settings: Claude Enterprise, where employees use the assistant in routine work, and Claude Platform, where developers manage API keys, workspaces, workspace membership and agent deployments.

Through Claude's Compliance API, Reco provides visibility into Claude Enterprise activity. Its integration with Claude Platform gives security teams a view into development settings, including users, groups, roles, permissions, projects, activity logs, workspaces, API keys, access and configurations.

Another part of the offering focuses on AI agents built and deployed on Anthropic's infrastructure. Reco maps each agent's model, version history, tools, permission policies and connected MCP servers, then relates that information to identities, permissions and links across more than 230 applications.

This broader context is central to the company's pitch. Rather than monitoring Claude as a standalone system, Reco correlates Claude activity with signals from across the enterprise, including applications, endpoint and network signals, to spot risks that might not appear in single-source monitoring.

Examples include an agent with broad access to sensitive data, an API key that remains active after its original purpose has ended, or an account that still has Claude access after an employee has changed roles or left the business.

Natural-language queries

The integration also lets security teams use Claude as an interface for investigations into Reco data. Through Reco's MCP server, teams can ask natural-language questions about access, ownership, unusual activity and risky connections, and receive answers based on information in the Reco Graph.

Findings from those investigations can then be routed into existing SIEM, SOAR and ticketing workflows for response and remediation. Reco presents this as a way to reduce the need for security analysts to switch between multiple tools while examining incidents or access issues.

Ofer Klein, Chief Executive Officer and Co-Founder of Reco, said Claude now plays a more central role in corporate technology environments.

"Claude is becoming part of the enterprise operating fabric, not just another AI tool," said Ofer Klein, Chief Executive Officer and Co-Founder of Reco.

"Security teams need to understand who is using it, what agents and applications it connects to, what permissions are involved, and what risk is created when AI activity moves across the business. Reco helps organizations govern Claude the same way they govern critical enterprise applications such as Okta, Salesforce and Microsoft 365," Klein said.

Broader AI oversight

The launch reflects a broader shift in enterprise security as companies try to apply more conventional governance methods to AI systems and the agents built around them. As AI assistants become embedded in internal operations, many security teams are concerned not only with model output or prompt behaviour, but also with access control, persistent credentials, connected tools and the movement of data between systems.

Reco is targeting six areas that have become central to security discussions around Claude: shadow AI and unauthorised usage, sensitive data leakage, prompt injection, API key exposure, excessive agent agency, and access control and offboarding gaps.

The Claude Security integration is available immediately through Reco and its partners worldwide. The company says its platform covers more than 230 applications and includes more than 1,000 detection controls.