Workday to Acquire Honeydew for a Shared AI Semantic Layer

Workday announced on November 18, at its Rising EMEA event in Barcelona, that it intends to acquire Honeydew, a semantic-layer company whose technology maps business definitions and applies the right one based on user, use case, and context. No purchase price was disclosed, and the deal is expected to close in Workday's fiscal Q4 2027 (ending January 31, 2027), subject to customary conditions. According to reporting on the acquisition, Honeydew's "semantic compiler" converts governed business definitions into queries that analytics tools and AI agents consume consistently, rather than letting each application interpret a metric like headcount or revenue on its own.
The problem this is solving is an old ontology problem wearing an AI badge
Strip away the AI framing and the pitch is familiar to anyone who has modeled a domain: different systems (HR, finance, a BI dashboard, a chatbot) each define "active employee" or "qualified pipeline" slightly differently, producing contradictory answers to what looks like the same question. Workday's stated plan is to combine Honeydew's compiler with Workday Data Cloud so that definitions are centralized once and consumed everywhere, instead of being re-implemented per application. That is a semantic layer in the literal sense, not a marketing one: a controlled vocabulary with relationships, bound to query logic, sitting between raw tables and the tools that read them.
Why vendors keep re-discovering this
Honeydew is not an isolated case; Neo4j, Databricks, Snowflake, and now Workday have each shipped their own business-ontology or semantic-compiler layer this year, usually justified by the same complaint — AI agents hallucinate or disagree with each other when the same metric has no single governed source of truth. What these products converge on, whether or not they use the word, is that an ontology-first data model is what keeps an AI agent's answer traceable back to a definition a human actually approved. Enterprises adopting Workday's new layer, or any of its competitors, still have to do the unglamorous part themselves: auditing which systems hold the authoritative version of each business term and wiring the integration so the semantic layer is actually fed clean, current data rather than another set of conflicting exports. Rolling that out across a Salesforce-centered sales stack alongside an HR and finance platform is exactly the kind of cross-system data-governance work that Salesforce enterprise integration projects are built to handle, since CRM data is usually the first place inconsistent business definitions surface.
What to watch next
Workday has not published a customer availability date for the integrated product, so the near-term signal to watch is whether early adopters report a drop in cross-tool metric disagreements once Honeydew's compiler sits in front of Workday Data Cloud, or whether the semantic layer becomes one more system that itself needs reconciling against the others.
Keep reading

Neo4j Context Engine ships a 6-layer Business Ontology for AI agents
Neo4j's Context Engine adds a structured business-ontology layer so AI agents reason over governed definitions.

NVIDIA Open-Sources Auto Ontology for AI Agents
NVIDIA's Auto Ontology builds semantic models from database metadata to help AI agents query enterprise data.