Enterprise AI can draft a policy, summarise a financial report and even recommend an action, but ask it to process a payroll exception or approve a role change and the question becomes less about capability and more about accountability.
Workday has moved to address that gap with a headless version of its Trusted Agent Platform. Announced in Hong Kong, the platform enables organisations to build, run and govern AI agents across HR and finance using their preferred development tools and models, while requiring every agent to operate under Workday’s permissions, approvals and audit trails.
The platform brings together Workday Data Cloud, Agent-Ready Tools, Developer Agent, Custom Agents and Agent Harness, all now generally available.
“Headless can’t mean lawless,” said Gabe Monroy, chief technology officer at Workday. “The rules for HR and finance already live in Workday, so customers don’t have to rebuild them for every agent they create.”
The significance for Asia-Pacific enterprises lies in the region’s increasingly active approach to AI governance. Hong Kong’s 2026 Policy Address sets out seven areas of AI risk governance, including safe AI applications, liability for AI-related incidents and safety guidelines for AI agents, with wider AI Agent safety-management guidelines planned for 2027. In Singapore, the Monetary Authority of Singapore has issued AI Risk Management Guidelines covering governance, lifecycle controls and third-party AI oversight for financial institutions.
Workday’s Agent Harness is designed to meet that operational reality. It supplies each agent with task-specific context, including the individual’s role, reporting line, access rights and applicable policies, then governs each action and routes tasks to an appropriate model based on cost and performance.
Agent-Ready Tools, built on the Model Context Protocol, provide approved routes for compatible agents to read Workday data and take action. Live Data Query allows teams and agents to query live Workday data in SQL through Workday Data Cloud, avoiding stale copies and fragile integrations.
“MCP gives agents a standard way to connect with the systems where work happens,” said Ankur Sinha, head of enterprise and verticals at Anthropic.
The platform also supports custom development. Teams can describe an agent in natural language using tools such as Claude Code, Codex or Cursor, then deploy it on Agent Harness. They can choose models from Anthropic, Google, Meta, Mistral, OpenAI and NVIDIA.
The practical implication is clear. AI agents will increasingly sit inside payroll, procurement, financial close and workforce processes—but their value will depend less on model choice than on whether every action can be traced to an authorised user, a valid policy and a complete audit trail. Workday’s platform is an attempt to make that standard the default, rather than an afterthought.
