As layoffs continue to rock the technology world, the need for technical talent in organizations is only increasing, putting a premium on how to manage in-house talent.
TechWolf, a startup based in Ghent, Belgium, has taken a unique approach to addressing this need: The company built an AI engine that ingests data from internal workflows and learns about the people performing those tasks. This data is then transformed into information that helps managers and internal recruiters better assess the interests and skills of different employees, connect them to different projects, and ultimately provide better training and more.
The company has been making a lot of buzz for its technology and boasts an impressive list of clients, including GSK, HSBC, and Booking.com, and now the company has raised around $43 million in funding ($42.75 million to be exact) to expand.
London-based Felix Capital is leading the Series B, with three of the HR industry giants — SAP, ServiceNow and Workday — co-investing for the first time. Other investors include Acadian Ventures, Fortino Capital Partners, Notion Capital, Semper Virens, 20VC and unnamed “AI leaders” DeepMind and Meta. As far as we understand, the startup is currently valued at around $150 million.
CEO Andreas de Neve co-founded TechWolf with Jeroen van Houtte and Michaël Wourneau in 2018 while the three were still computer science students at Ghent University in Belgium and Cambridge University in the UK.
The original plan, he said, was for the startup to build an HR platform by building its own language model “like ChatGPT” to help with external talent discovery and hiring.
“It was a failure,” he said simply. Hiring, or at least the part they were trying to address, wasn't so much broken: Employers “didn't need AI to sift through the good applicants from the bad.”
However, the founders realized that their target customers had a different problem to solve.
“They said, 'Hey, is there a chance we could use this AI model on our 40,000 employees, rather than on applicants, because there might be people we could hire internally,'” De Neuve says. “The HR leaders steered us towards the right problem to solve, which was identifying employee skills.”
“What do you actually do?” is a question often asked as a joke about Chandler (the IT worker) from the TV show Friends. But it's a huge problem in real-world business, and it gets worse the larger the organization. “You might have 100,000 employees, all of whom are highly competent and spend a lot of their time in software systems that create data,” De Neve says. “But structurally, these companies know very little about these employees. And that's what we set out to do.”
That's exactly the kind of problem AI can solve, he says. “We've started building language models that integrate with the systems people use at work – project trackers, documentation systems for developers, research repositories for researchers – and then from all that data we infer what skills workers have. You can think of it as a set of AI models connected to the digital exhaust of an organization.”
TechWolf touches on several key trends worth noting in the market today.
What is the real Innovator's Dilemma? The seminal book “The Innovator's Dilemma” compellingly portrays how even the most successful large companies can be disrupted by smaller companies that move to respond nimbly to change. But looking at this from another angle, the core asset that helps some organizations work more flexibly than others is their people. The ability to organize teams around different projects and goals will likely make or break those efforts. And it turns out organizations are willing to pay big bucks for technology to help them with that task. LLM vs MLM vs SLM. “Large” language models and the companies building them continue to attract a lot of interest. And the key word here is “generative” because they are the foundation of popular generative AI applications like ChatGPT, Stable Diffusion, Claude, Suno, and more. But there is no doubt that there is growing demand for “smaller” language models that can be applied to very specific use cases. These models are less complex to build and operate, and ultimately more likely to be less constrained and less hallucinatory. TechWolf isn't the only company operating in this space, nor is it the only one garnering investor attention. (Another example is the startup Poolside, which also builds AI for a specific use case: developers and their coding tasks.) Narrowing the focus is indeed crucial. I asked De Neve whether TechWolf had ambitions to leverage the platform to expand into other areas, such as enterprise search or business intelligence. After all, the company is already ingesting so much enterprise information, so wouldn't it be an even easier step to build more products around it?
No, says De Neuve emphatically. “We have data processing capabilities that no one else in the market has, but we are very focused on solving the skills problem because there is already so much demand for us in the markets where we currently operate.”
At a time when the AI world can feel quite noisy, focus rings like a clear bell and may be one of the reasons investors are taking an interest in companies like this.
Julian Codorniou, the Felix partner who led the deal, believes Techwolf can out-do even bigger players in other areas, such as AI-based enterprise search. “If you do one thing well, you really deliver,” he said. “Techwolf doesn't want to be Workday or ServiceNow. We want to be the Switzerland of HR.”