BARCELONA, November 19, 2026 — Workday announced its intent to acquire Honeydew, a semantic layer for AI and business intelligence, at Workday Rising EMEA on November 18. The deal targets a growing inconsistency problem: dashboards, chatbots, and AI agents often return different answers for the same metric because each tool applies its own definition.
The Inconsistency Problem: Different Tools, Different Answers
Ask three systems for a company’s headcount and you get three numbers. One counts only full-time employees, another includes contractors, a third adds approved-but-unfilled roles. None is technically wrong, but the inconsistency breaks trust in AI-driven decisions. Workday CTO Gabe Monroy put it bluntly: “An AI model shouldn’t be improvising the calculation behind a payroll number.” Without a shared definition layer, data teams spend months writing custom code to intercept LLM requests, hardcoding business rules into brittle pipelines, and stuffing massive data dictionaries into every prompt. When teams switch models or platforms, they start over.
How a Semantic Layer Maps Business Terms to Governed Queries
A semantic layer sits between business users and raw data, mapping familiar terms like “revenue” or “headcount” to precise, governed queries. The concept originated with Business Objects, which secured U.S. Patent 5,555,403 in September 1996 for letting end users query relational databases without knowing SQL. Modern semantic layers use standards such as RDF, SKOS, and OWL, growing alongside knowledge graphs and retrieval-augmented generation (RAG) as infrastructure for enterprise AI. Honeydew’s semantic compiler turns business definitions into deterministic queries, natively capturing multiple definitions for context-specific use — so the right definition applies based on who is asking, the use case, and the situation.

Honeydew’s Semantic Compiler and Live-Data Approach
Honeydew was founded four years ago on the bet that companies need a single source of truth for meaning, eliminating dozens of scattered definitions for metrics like revenue. Its approach is SQL-based and centered on the data warehouse, but with a key difference: it applies definitions directly to live data rather than extracted copies. Most semantic layers rely on data pulled into a warehouse or lake, leaving answers dependent on the last refresh. Honeydew works on live data, so definitions stay current. As Honeydew CEO David Krakov said, “Getting AI to write database queries was never the hard part. Getting it to understand the true definitions behind your company’s data is.”
Workday Data Cloud Integration and Agentic AI Context
Workday plans to embed Honeydew into Workday Data Cloud, giving agents governed business context to act quickly and consistently. The integration moves business logic out of the prompt and compiles it into deterministic queries before a request reaches the model. Permissions travel with the definitions, so each person or agent sees only what they are allowed to see, and invalid or unauthorized requests are blocked before any query runs. Yasmeen Ahmad, SVP and GM of Workday Data Cloud, framed the shift: “The greatest bottleneck in enterprise AI today isn’t reasoning, it’s context.” The transaction is expected to close in the fourth quarter of Workday’s fiscal year 2027, ending January 31, 2027. Financial terms were not disclosed.
Industry Shift: Spending Moves from Models to Data Infrastructure
The acquisition signals enterprise AI spending reaching beyond models to the data layer that agents depend on. Workday describes itself as the trusted agentic platform for HR, finance, and IT, used by more than 11,500 organizations worldwide, including more than 65% of the Fortune 500. By compiling definitions into governed queries upfront, Workday says customers can expect more accurate and efficient answers, fewer tokens, and lower costs. The shift underneath is clear: the competitive edge is moving from who has the best model to who has the best governed context.
Historical Evolution from Business Objects to AI-Era Knowledge Graphs
The semantic layer has traveled a long road from Business Objects’ 1996 patent for shielding end users from SQL. In the late 2010s, independent products emerged — universal semantic layers, metrics layers, headless BI — that define metrics in a central repository outside any single analytics tool. Honeydew’s compiler represents the next step: a live, context-aware layer that serves both human analysts and autonomous agents from the same governed definitions. The deal closes before the end of January 2027.
Sources
- Workday Inc.: Workday Announces Intent to Acquire Honeydew, a Semantic Layer for AI and Business Intelligence
- Semantic layer
- Workday to acquire Honeydew, a semantic layer for AI and BI
- Semantic caching helps enforce AI agent governance inline – SiliconANGLE
- Workday to buy Honeydew, a semantic layer for AI agents
- Workday Rising EMEA – Workday acquires Honeydew for semantic consistency, strengthens EU Sovereign Cloud
Photo: Stephen Dawson / Unsplash (Unsplash License)





