Opinion A $4.99 App Now Prescribes Acne Drugs in Utah With No Doctor in the Room — and the Oversight Ladder After 500 Patients Is the Real Story News AMD and 14 Korean AI Chip Firms Will Build a Heterogeneous Compute Reference Model Pairing Instinct GPUs With Homegrown NPUs Voice & Audio ElevenLabs Commits 'Hundreds of Millions of Dollars' to India — Its Biggest Market by Signups, With 100 Million AI Agent Conversations a Year Business SAP Acquires TechWolf, the Belgian Startup Whose 'Context Graph for Work' Tells Its HR Agents What Every Employee Actually Does Business Anthropic's IPO Prospectus Puts Dario Amodei's 2025 Pay at $18 Million — Mid-Pack Among Big-Tech CEOs Productivity Atlassian Puts GPT-6 Across Jira, Confluence and Rovo as Its 3,000-Developer Codex Bet Goes Company-Wide Security Hadrian Raises $40 Million to Stress-Test the New Attack Surface: AI Agents That Can Be Talked Into Leaking Data Business Vinci Raises $250 Million at $1.5 Billion to Automate the Physics Behind Chip Design Opinion A $4.99 App Now Prescribes Acne Drugs in Utah With No Doctor in the Room — and the Oversight Ladder After 500 Patients Is the Real Story News AMD and 14 Korean AI Chip Firms Will Build a Heterogeneous Compute Reference Model Pairing Instinct GPUs With Homegrown NPUs Voice & Audio ElevenLabs Commits 'Hundreds of Millions of Dollars' to India — Its Biggest Market by Signups, With 100 Million AI Agent Conversations a Year Business SAP Acquires TechWolf, the Belgian Startup Whose 'Context Graph for Work' Tells Its HR Agents What Every Employee Actually Does Business Anthropic's IPO Prospectus Puts Dario Amodei's 2025 Pay at $18 Million — Mid-Pack Among Big-Tech CEOs Productivity Atlassian Puts GPT-6 Across Jira, Confluence and Rovo as Its 3,000-Developer Codex Bet Goes Company-Wide Security Hadrian Raises $40 Million to Stress-Test the New Attack Surface: AI Agents That Can Be Talked Into Leaking Data Business Vinci Raises $250 Million at $1.5 Billion to Automate the Physics Behind Chip Design

SAP Acquires TechWolf, the Belgian Startup Whose 'Context Graph for Work' Tells Its HR Agents What Every Employee Actually Does

SAP Acquires TechWolf, the Belgian Startup Whose 'Context Graph for Work' Tells Its HR Agents What Every Employee Actually Does

SAP has agreed to acquire TechWolf, the Ghent-based AI company that infers employees' skills and tasks from the HR and business systems they already use, for an undisclosed sum. TechWolf becomes the intelligence core of SAP SuccessFactors, will keep serving non-SAP customers under CEO Andreas De Neve, and the deal is expected to close in Q4 2026 pending regulatory approval.

SAP said on Tuesday it has agreed to acquire TechWolf, a Belgian AI company that maps what employees work on and which skills they use, for an undisclosed sum. The deal, announced at SAP's Connect conference in Las Vegas, is expected to close in the fourth quarter of 2026 subject to regulatory approval, and hands the German software giant something its own HR software has never had: a continuously updated picture of what its customers' workforces actually do, rather than what their self-assessments claim.

TechWolf, founded in Ghent in 2018 by Andreas De Neve, Jeroen Van Hautte and Mikaël Wornoo, plugs into the HR and business applications a company already runs and infers skills from people's actual work. It calls the result a "context graph for work," and it covers three layers: the work itself, broken down to the tasks inside each job; the skills employees put to use; and the external labor market. The platform then lines that organizational model up against where the business says it is heading, giving managers deciding on hiring, retraining or restructuring something firmer to stand on than an annual skills survey.

For SAP, the acquisition is less about the product than about what its AI agents are standing on. Manoj Swaminathan, president and chief product officer for SAP Autonomous Suite, said TechWolf's graph provides "an excellent grounding layer for agent queries regarding work and skills planning and talent management," and — in a detail that says a lot about how enterprise AI is now sold — that it "makes token usage more efficient," cutting the cost of running HR agents. The company plans to make TechWolf the intelligence core of SuccessFactors, feeding skills and work data into tools for workforce planning and organizational redesign, including a Workforce Planning Assistant it introduced at Connect for early adopters this quarter.

The deal fits a pattern that has crystallized across enterprise software this year: incumbents are buying the data layer an AI agent needs to be trustworthy rather than waiting to be disrupted. SAP bought Reltio earlier in 2026 to give Joule's agents one consistent view of customers, suppliers and products; Atlassian made the same argument for its Teamwork Graph in May, claiming agents with access to its context used 48 percent fewer tokens. TechWolf was already an SAP partner — SAP was among the investors in its $42.75 million Series B in 2024, a round led by Felix Capital that also included ServiceNow and Workday, two of SAP's direct competitors for HR software budgets.

TechWolf is not a small operation hiding inside a big logo. Its customers include Booking.com, HSBC, MetLife, PayPal, AMD, Ericsson and GSK, and it will keep serving non-SAP customers after the acquisition closes. The company stays independent under co-founder Andreas De Neve, keeps its Ghent headquarters, and is familiar to developers for the JobBERT models it publishes openly on Hugging Face, which map job titles and descriptions into a shared vector space — the datasets alone have racked up millions of downloads.

The broader read: as companies reorganize around AI, the question "which skills do we have and which are we losing?" has stopped being an HR reporting problem and become an input to agent-driven planning. HR agents are only as useful as the workforce context they can draw on, and SAP has decided that buying that grounding layer beats building it. Once the deal closes, the contest shifts from who has the best model to who has the most accurate, most current map of the work itself.

Comments (0)

Log in to join the discussion

Log In

No comments yet