Experience the difference. Meet your peers, talk with experts, build your network, and turn insights into action.
Atul Tulshibagwale
Senior Director, Continuous Identity Strategy
CrowdStrike
Speaker Bio
Atul is a Board Member of the OpenID Foundation and the Senior Director, Continuous Identity Strategy at CrowdStrike. He was previously the CTO of SGNL, which CrowdStrike acquired in February 2026. Named in the Okta “Identity 25”, Atul is a federated identity pioneer and the inventor of the Continuous Access Evaluation Protocol (CAEP). He was previously at Google, where his seminal blog post kicked-off the industry-wide movement that culminated in the OpenID Shared Signals working group, which he co-chairs. His leadership in developing and promoting SSF and CAEP, the critical zero-trust standards, has been influential in their widespread adoption. Apple, Okta, Cisco and others have announced support for these standards. Previously, Atul was a co-founder and the CEO of Trustgenix, a federated identity pioneer that was acquired by HP. Trustgenix contributed to the development of federated identity standards such as SAML 2.0 and the Liberty Alliance Framework.
Show More
Show Less
Session
Tuesday, December 08, 2026 12:00 PM - 12:30 PM PST
Executive Story Panel: OpenID: Securing AI Agents with Existing IAM Standards
Atul Tulshibagwale,
Senior Director, Continuous Identity Strategy, CrowdStrike
As enterprises race to adopt AI, identity and security practitioners face a familiar set of problems in an unfamiliar context: discovering which agents are running (and preventing "shadow AI"), ensuring only approved agents and LLMs are used, enforcing fine-grained access to specific resources, revoking access when conditions change, and auditing every action across call chains that traverse systems and organizations. The Model Context Protocol (MCP) has emerged as the de facto communications layer connecting agents to enterprise systems, and it supports OAuth, but OAuth may not provide all the answers. While some might be inclined to invent new standards to address these gaps, we propose achieving this using existing standards. The building blocks already exist. In this session, we'll walk through five concrete best practices, each grounded in an existing or emerging open standard, that together tame AI's unpredictability with deterministic controls.
...
Show More
Show Less