“I’ve never believed that learning is just about logic and rules. I wanted to truly understand how computers could work with language and learn its subtleties.”
When it comes to Artificial Intelligence, Jan Scholtes is a true pioneer. Back in 1985, he was already working on AI applications in Delft, even though, as he puts it, “it didn’t really work back then.” Fast forward nearly forty years, Scholtes is now a full Professor holding the extra-ordinairy chair of Text-Mining at Maastricht University, advising companies on AI implementation and recently taking on the role of Venture Partner for AI at Endeit Capital. Together with Endeit’s partner Jelle-Jan Bruinsma, we explore Scholtes’ journey through the evolving world of AI and how his dual expertise as a scientist and entrepreneur strengthens Endeit’s investment strategies.
Scholtes began his career as a math and computer science student in Delft, but his interests extended beyond just those fields. “At TU Delft, human language was viewed as a kind of programming language, but that didn’t resonate with me,” Scholtes explains. “I’ve never believed that learning is just about logic and rules. I wanted to truly understand how computers could work with language and learn its subtleties. So, I decided to pursue studies at the Department of Computational Linguistics at the University of Amsterdam.” This pivotal move took him to the P.C. Hoofthuis, where, in the early 1990s, he conducted PhD research on the use of neural networks for Natural Language Processing as well as their application in the first generation of search engines.
“I’ve never believed that learning is just about logic and rules. I wanted to truly understand how computers could work with language and learn its subtleties.”
Scholtes’ academic achievements laid the groundwork for what would eventually become an international business. His specialization in search technology and text analytics caught the attention of law enforcement agencies, war tribunals, and R&D departments of global corporations. In 2002, he acquired the American company ZyLAB, transforming it into a globally successful enterprise. “We sold our advanced search and analytics systems in over 50 countries to organizations ranging from the White House to parliamentary inquiry commissions in the Netherlands,” he recalls. What set us apart was our ability to support hundreds of languages and handle all kinds of digital documents that you’d encounter in real-world scenarios.” Over time, ZyLAB shifted its focus from public to private sectors. In 2021, the company was successfully sold to American investors.
“I was appointed an extra-ordinary (full) Professor at Maastricht University in 2008, dedicating one day a week to teaching and research. After selling ZyLAB, I had more time for academia,” says Scholtes. Next to his courses Information Retrieval and Text-Mining he also became responsible for lectures in Advanced Natural Language Processing (NLP). “By then, I was already in regular contact with Endeit,” he shares. The partnership became more formal with Endeit’s portfolio company IPRally, which immediately piqued Scholtes’ interest. “It reminded me of the approach we used at ZyLAB.” Bruinsma adds, “We saw potential but needed Jan’s expertise to assess the technical depth. The key question was: Does IPRally have something unique? Jan’s analysis confirmed they had solved a complex problem in a large market.”
The collaboration between Jan and Jelle-Jan proved so valuable that creating an official role as Venture Partner for AI was the natural next step. “At Endeit, we already had several venture partners across different sectors, but we didn’t have one for AI,” says Bruinsma. “If anyone truly understands how AI works, it’s Jan.” Scholtes’ expertise has a direct impact on Endeit’s investment decisions. “We focus on sectors like Fintech, Climate Tech, Supply Chain and Industrial Automation, Sales Tech & Future of Work. We get excited when we see business models where AI is implemented in the right way.”
“If anyone truly understands how AI works, it’s Jan.”
Scholtes’ decades of experience give him a keen insight for spotting genuine innovation in AI. “With AI, things either go very well or very wrong, very quickly. There’s a huge marketing hype, but if you understand the underlying math, you know what’s possible and what’s not. Take ChatGPT, for example—it’s a language model, and you should only use it for tasks suited to that kind of technology,” Scholtes notes.
The evolution of AI applications is fascinating, he observes. “In the past, searching was about finding the needle in a haystack. Text-mining advanced to questions like: What does the needle look like, and where is the haystack? Now, we have conversational AI, where people interact directly with language models to find information. The chatbot searches for them and presents the results as a dialogue.”
However, Scholtes stresses the importance of checking for errors in AI systems. “There are critical challenges in using AI for legal or medical applications. If you were to generate summaries or have a conversation in those fields using ChatGPT, mistakes happen too often. We need to understand why and detect these errors—or better yet, prevent them. Right now, not enough companies are addressing this properly, often because they lack knowledge of the right techniques or access to quality data. That’s where we can guide them.”
“With AI, things either go very well or very wrong, very quickly. There’s a huge marketing hype, but if you understand the underlying math, you know what’s possible and what’s not.”
“The barriers to entry are dropping,” observes Bruinsma. “This changes how we approach investments and what we expect in the long term.” Scholtes adds, “Startups don’t always see the full potential of what they’re (often intuitively) doing well. We can help them with positioning and ensure they apply AI responsibly and legally across broad contexts. Together, we can build truly strong business cases and maximize their potential.”
Endeit refers to the following statement in connection with the sustainable finance disclosure regulation (SFDR), available here.
Get in touch with us by submitting your plan or by contacting one of our team members directly.
If you want to receive our newsletter, please put ‘newsletter‘ in the additional message box. ⮕