Linksight: Shared insights from encrypted data

This article was originally published in Dutch in issue #10 of TTT magazine by Novel-T. For your convenience, we have translated it here.

Insights from combined data are important for better healthcare, smarter business operations, and more effective policy. In practice, however, you cannot simply bring datasets together, because of privacy and data legislation and the increasing risk of data breaches. So how do you still extract new insights from combined data? Linksight, a spin-off of TNO, is developing a way to gain insights from encrypted data with the help of TTT-AI. CEO Martine van de Gaar talks about their innovation.

New analyses

Through Linksight, organizations have access to insights from various datasets without ever seeing the data itself. “Every piece of data is encrypted,” Martine explains. “So you don’t know what it is, but you can still compute with it, which means you can extract the insights. In effect, you can combine datasets without sharing sensitive data. That makes new analyses possible.” The technology can be useful, for example, in tackling waiting list problems in mental healthcare. “Those lists can be compared with one another to see how many unique people appear on them. Perhaps a person registered twice, or with multiple organizations. You can find that out without seeing the individual data.”

How can you learn something about something you have never seen? That is precisely what Linksight makes possible, Martine explains. “Our technological breakthrough is that we have translated a cryptographic method into workable protocols, supported by data collaboration governance. The result is a product that lets end users extract insights from combined, encrypted data. The data stays at the source, but the encrypted information can still be used for computation. So you get the insights, but not the data itself.” This opens up many new possibilities for organizations. “You have control over your own data,” Martine says. “You decide who is allowed to do something with your data — or with part of your data. And you can decide that anew for every collaboration.”

Martine van de Gaar – CEO Linksight

Product introductions

Most of Linksight’s clients are in healthcare, a sector that works extensively with sensitive data. “We carry out studies into effectiveness, such as process improvements or product innovations,” Martine says. “Think, for example, of eye-drop glasses that could make it possible to deploy less home care after cataract surgery. You then have to investigate whether that is actually the case: can people use them independently? And are enough drops being applied? For that, you need data from various parties: from the hospital you need to know who was operated on, from pharmacists who they handed the glasses out to, from home care who they provide care to, and so on. So you have different parties with data that you cannot simply throw onto one pile, while you do want those insights in order to improve healthcare. These are the kinds of things we can calculate, to see whether such a product introduction really helps organizations move forward.”

“I can also picture a future where pooling datasets onto one pile will seem as strange to us as smoking in a café. That shift in mindset is something I really want to make possible.”

— Martine van de Gaar, CEO Linksight

Self-sovereignty

With this technology, Linksight meets a major need of organizations in 2026: self-sovereignty — in other words, control over your own data. “We give our clients the opportunity to make an efficiency gain without having to compromise on privacy,” Martine says. “With the rise of AI, more and more models are becoming available. Our solution makes it possible to train all those AI models in a secure way. This also makes it possible to use sensitive data for AI models. I can also picture our technology becoming completely mainstream and bringing about a different mindset. That pooling datasets onto one pile will seem as strange to us as smoking in a café. That shift in mindset is something I really want to make possible.”

Linksight received support from TTT-AI. “LUMO Labs was very helpful in that,” Martine says. “As former entrepreneurs, they know the ropes. They know what phase you are going through and what challenges come with it. It was great to have them as a sounding board, a sparring partner, and added expertise.”

Source: TTT magazine #10, Novel-T