Workday adds tools to build and verify AI agents

Workday Inc. on Thursday announced new capabilities for building and verifying AI agents on top of its platform, expanding Workday Build with tools aimed at developers working in enterprise environments. The company unveiled the features during DevCon 2026, its annual developer conference.

The update includes Developer Agent, Agent-Ready Tools and a new security framework called Agent Passport. Workday said Developer Agent is designed to fit into the tools developers already use, including Cline, Claude Code, Cursor and Google Antigravity, so they can move from simple requests to working apps or agents more quickly.

According to the company, Developer Agent can draw from a library of more than 50 reusable agent skills built on the AgentSkills standard. Those skills are meant to help developers add enterprise knowledge to agents without starting from scratch each time.

Workday also announced Agent-Ready Tools, a set of connectors that expose agents through open protocols including Model Context Protocol and Agent to Agent. The company said that when an external system uses these tools, its actions inherit Workday’s data governance, security frameworks and business processes.

Dean Arnold, vice president of AI platform at Workday, told SiliconANGLE that the goal is to let customers use Workday’s tools in many different platforms, not only inside Workday itself. He said the company wants developers to be able to bring enterprise controls into outside editors, tools and coding agents.

Workday framed the release as its largest update so far to Workday Build since the company added low-code and no-code development tools. Jay Wieczorkowski, vice president and general manager of the developer platform, told SiliconANGLE that agentic AI is changing how developers work and that they now have to maintain “enterprise truth” inside the agents they build.

That focus is important in HR, finance and payroll systems, where Workday said agents must preserve business logic, rules, user roles, company context and compliance limits. Wieczorkowski said there is “zero room for error” in areas such as payroll, benefits and the ledger.

Workday said Agent Passport is meant to add a verification layer before an agent reaches production. The framework shows which security and compliance tests an agent has passed, who verified it and which standards were used.

The company said this should help security and audit teams quickly identify which agents are current and approved for production use. Workday already offers governance controls and guardrails for agents in production, but Agent Passport adds an independent verification step.

Arnold described the process as “red-teaming up front,” saying the analysis happens before shipment and then stays current. He said Cisco Systems Inc. is a launch partner and is helping provide the initial “stamps” that will appear on agents as part of Passport.

Workday said the broader goal is to let developers choose how they use AI coding agents while keeping enterprise data, policies and compliance controls in place. Wieczorkowski said the new tools do not remove developers from the process; they make them more effective while still keeping them in the loop.