Workday plans to buy Honeydew, a start-up that tells AI agents and business-intelligence tools how a company defines its numbers. Workday announced its intent to acquire Honeydew on Wednesday at its Workday Rising EMEA event in Barcelona and did not disclose terms.
Our number: Workday expects the deal to close by 31 January 2027, the end of its fiscal year, which is within 74 days of the announcement, by our arithmetic.
Business Pill · Least Privilege
An evergreen Business Engineer explainer on least privilege and access control. It teaches the general idea only and does not discuss this deal.
The key insight: Workday intends to buy Honeydew, a semantic layer that gives AI agents and BI tools the same governed business definitions; it did not disclose terms and expects to close by 31 January 2027, within 74 days by our arithmetic.
What Workday Announced
Workday describes Honeydew as “a semantic layer for AI and business intelligence (BI)”. It says Honeydew’s semantic compiler, which “turns business definitions into precise governed queries”, will work with Workday Data Cloud.
The release gives an example: a dashboard, a chat tool and an AI agent might each answer a headcount question differently, depending on whether they count contractors or unfilled roles. Workday says Honeydew “pulls the exact right definition based on the person or agent asking, the specific use case, and the situational context”.
Gabe Monroy, Workday’s chief technology officer, said: “An AI model shouldn’t be improvising the calculation behind a payroll number”. He said that with Honeydew, “the model says what it needs and Workday builds the exact query, the same way every time, within each person’s or agent’s permissions”.

Why Definitions Matter for Agents
Workday says that today data teams resort to “schema-stuffing”, packing data dictionaries, join rules and definitions into every AI prompt. It says Honeydew will move business logic out of the prompt and compile it into queries before a request reaches the model, and that customers could expect “fewer tokens, and lower costs”.
David Krakov, Honeydew’s chief executive, said: “Getting AI to write database queries was never the hard part”. Yasmeen Ahmad, who runs Workday Data Cloud, said: “The greatest bottleneck in enterprise AI today isn’t reasoning, it’s context”.
Workday says more than 11,500 organisations use it, including more than 65% of the Fortune 500, which is more than 325 companies, by our arithmetic. Workday’s own systems were in the news this week for a different reason: an 18-hour outage.
The Structural Read
As we read the release, Workday is buying the layer between the model and the data: it says business logic will be compiled into queries before a request reaches the model, rather than packed into prompts.
As we read Monroy’s remarks, the pitch is consistency under permissions: Workday says the same query is built every time, within each person’s or agent’s permissions.
Gabe Monroy, Chief Technology Officer, Workday
“An AI model shouldn’t be improvising the calculation behind a payroll number”
Three Implications
FOR WORKDAY CUSTOMERS Workday says Honeydew would apply a customer’s definitions directly to live data in Workday Data Cloud.
FOR DATA TEAMS Workday says data teams spend months writing custom code and hardcoding business rules to connect AI agents to enterprise data.
FOR AI COSTS Workday says customers could expect greater accuracy and efficiency, fewer tokens and lower costs.
The Business Engineer Lens
This story maps onto the Business Engineer framework Enterprise AI: From Software to Substrate.
The analysis puts it this way: “The value proposition shifts from features to integration depth.”
As we read Workday’s release, Honeydew is integration depth for agents: the definitions behind each number would live inside Workday rather than in each tool or prompt. The analysis predates this deal and does not discuss it.
What Is Not Established
Workday does not disclose the price, Honeydew’s revenue, customers or headcount, or how many Workday customers would get Honeydew’s features or at what cost. The deal has not closed and is subject to customary closing conditions.
The token and cost savings are Workday’s expectation, not a measured result. We did not contact Workday or Honeydew.
The Bottom Line
Workday intends to buy Honeydew so that AI agents and BI tools use the same governed business definitions. It did not disclose terms and expects to close by 31 January 2027, within 74 days, by our arithmetic.
97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Every fact about the deal, Honeydew’s product and the executives’ quotes is from Workday’s press release of 18 November 2026, read in full. The 74-day window and the 325-company figure are our arithmetic from Workday’s own figures. We did not contact Workday or Honeydew.
Sources: Workday press release, 18 Nov 2026









