Workday and Anthropic announced a partnership on October 13 to connect Claude with Sana, Workday’s agent experience and orchestration layer. The integration uses the Model Context Protocol to exchange tasks and context between the two environments.
According to Workday, employees working in Claude can use Sana agents to analyze permitted HR and finance information and take action. Examples include reviewing hiring plans or checking invoice exceptions. Existing Workday permissions and audit controls apply, and users confirm changes.
In Sana, longer processes remain anchored in Workday. Claude can assist with individual steps using the context needed for that step, while exceptions go to people for review. These are capabilities described by the companies, rather than independently verified deployment outcomes.
The Sana in Claude plugin is available to early access customers, with general availability expected in late 2026. Workday Agent-Ready Tools are already generally available through Workday Extend AI; access through the plugin is planned for later this year.
The partnership highlights a practical requirement for enterprise agents: reasoning must operate within business permissions and approval processes. Its usefulness will depend on reliable execution, clear exception handling and records that allow organizations to review what happened.