Workday, an enterprise AI platform for HR, finance and IT, has introduced Agent Passport, a system designed to test and verify AI agents, whether built by Workday or by third parties, before they go into production, and to continue monitoring them afterward. Each attestation under the program is tied to a public industry standard, including OWASP LLM Top 10, NIST AI RMF and MITRE ATLAS, giving security teams a signed, auditable record of what each agent has been tested for and which party performed the testing.
Agent Passport tests agents against risks including prompt injection, jailbreak and goal hijacking attempts, system prompt extraction, leaks of employee data, and unsafe outputs. Each result is signed by the partner that conducted the test, which Workday said makes records comparable across agents built by different vendors. Once an agent attempts to execute a task, Agent Passport monitors the action in real time and can allow, block or reroute it. If an issue is identified, a single revocation can stop, limit or restrict the affected agent according to company policy.
Dean Arnold, vice president of AI Platform at Workday, described what is at stake as agents take on more enterprise work:
"AI agents are now doing the most sensitive work in the enterprise, from onboarding employees to processing payments, and one insecure agent can leak employee data, break compliance, and put the company on the front page for the wrong reasons. Agent Passport gives companies confidence that every agent has been independently tested and verified, and the power to shut any of them down across the business the moment something changes."
Workday said most platforms that offer agent security testing perform that testing themselves, meaning a "safe" label often comes from the same vendor that built the agent. Agent Passport is structured differently, organizing each agent's record into three layers. The first covers broad trust areas that Workday defines and maintains, such as protection against attacks, safe behavior at runtime, and human oversight. The second layer consists of specific, testable claims tied to public standards, such as resistance to known attack techniques. The third layer holds the signed results from the partner that performed the testing, beginning with Cisco.
Because each check is tied to a public standard, Workday said security teams can compare agents from different vendors using the same criteria. If two agents carry the same check from two different partners, the company said, both were evaluated against the same bar.
Cisco, a US-based enterprise networking, security and infrastructure company, is the launch partner for Agent Passport, contributing its Cisco AI Defense product to independently test AI agents running on Workday before deployment and to continue protecting them at runtime against prompt injection, data leakage, jailbreaks and unsafe actions.
According to Workday and Cisco, Cisco AI Defense confirms that an agent resists attempts to override its instructions, prevents its own instructions from being exposed, protects sensitive employee information from leaking, and blocks harmful or policy-violating responses before they reach a user. The companies said these checks apply to any agent but carry particular weight for agents handling payroll, benefits and financial data.
DJ Sampath, senior vice president and general manager of AI Software and Platform at Cisco, framed the partnership around the scale at which agents are being deployed:
"Agents are going to be everywhere in the enterprise, and that only works if security teams have a clear, signed record of what each one has been tested for. Cisco AI Defense was built for exactly this kind of validation, and we're excited to partner with Workday to secure the agentic workforce."
Agent Passport will be available to early access customers in the second half of 2026, with general availability projected before the end of the year. The partnership between Workday and Cisco is active immediately, with additional joint capabilities expected to roll out over the coming quarters.
