Workday Acquires Honeydew to Govern AI Data
The deal adds a semantic layer that applies business definitions and permissions to live data within Workday Data Cloud.
Updated on Nov. 18, 2026 in Artificial Intelligence

Workday has acquired Honeydew, a semantic layer for AI and business intelligence, announced on November 18, 2026. Honeydew applies customer business definitions directly to live data in Workday Data Cloud and adds governance when a question becomes a query.
Why it matters
Most semantic layers operate on data already extracted into a warehouse or lake, so their answers depend on the last refresh. Keeping governance attached to live queries addresses how enterprise AI agents access business data.
Honeydew compiles business logic into deterministic queries rather than leaving it in the prompt, and Workday builds each query within the permissions of the requesting person or agent. Invalid or unauthorized requests are blocked before any query runs.
The players
Workday
Workday is an enterprise software company that provides human capital management and financial management applications for large organizations.
Honeydew
Honeydew is a semantic layer company that provides business intelligence and AI data definition tooling.
The details
Honeydew moves business logic out of the prompt and into deterministic queries, while Workday permissions travel with its definitions. The combination is intended to make agent responses as governed as the actions agents take, alongside Workday's agent governance.
Timeline
November 18, 2026: Workday announced the acquisition of Honeydew.
The Tech Race
The acquisition places Workday alongside enterprise data vendors competing to make AI agents reliable on business data, shifting the emphasis from periodically extracted warehouse snapshots to governed live queries. It reflects a broader push by enterprise software firms to build permission-aware AI infrastructure.
For employees and customers using Workday, the acquisition signals that AI-driven analytics will be tied to existing role-based permissions rather than to a separately maintained data copy. Organizations evaluating AI agents on business data may see governance features become a more standard part of these tools.
The takeaway
Governance for AI works best when it is enforced at the moment a question becomes a query, not only when data is loaded. Organizations should check whether their AI tools apply the same permissions as their underlying business systems.
Further reading
For more on how companies are putting guardrails on enterprise AI, see Artificial Intelligence.






