Okta Unveils Cross App Access to Secure Enterprise AI Agents

Okta is rolling out a new protocol designed to secure AI agents, with the goal of enhancing visibility, control, and governance over agent-generated interactions and cross-application communications. The initiative addresses emerging security challenges in enterprise environments driven by the growing adoption of AI tools.

Cross App Access was developed by Okta in direct response to the challenges associated with integrating AI agents with critical enterprise data and applications. “As we work actively with the MCP and A2A communities to improve agent functionality, their increased access to data and the explosion of cross-application connections are creating new identity security challenges. With Cross App Access, Okta is excited to provide oversight and control over how agents interact across the enterprise,” says Arnab Bose, CPO, Okta, to Security Boulevard.

In the modern enterprise security landscape, application-to-application (A2A) connections often operate without centralized oversight. This lack of visibility forces IT and security teams to rely on inconsistent methods for monitoring, reporting, and observing interactions, resulting in significant blind spots and the expansion of unmanaged perimeters, reports Security Boulevard.

The emergence of AI agents introduces non-deterministic access patterns, system boundary crossings, autonomous action triggers, and interactions with sensitive data, intensifying this security challenge. Up to 82% of executives see an urgent need to implement a robust AI governance solution to address this problem, reports Zogby Analytics. 

Cross App Access Capabilities

The Cross App Access platform is designed to enable independent software vendors (ISVs) to offer secure, enterprise-grade integrations for AI agents and other autonomous systems, including workflow automation tools. By shifting access control to the identity provider, such as Okta, ISVs can reduce security risks, simplify integration complexity, and strengthen support for compliance and governance requirements.

Historically, integrating AI tools with existing data systems has posed considerable obstacles, with many companies relying on ad hoc methods such as long-lived tokens and fragmented access controls. These practices increase risk and have slowed AI adoption for cross-application data access. Cross App Access is expected to allow companies to enhance both security and usability, empowering IT departments to manage agent access while delivering low-friction user experiences, says Okta.

This functionality supports secure interoperability between applications and AI systems, facilitating the adoption of innovative ISV solutions without compromising oversight or performance, says the company. The platform seeks to standardize and secure a rapidly growing area of vulnerability in modern IT infrastructure.