As part of the project, the team behind the start-up is developing an AI platform that will be able to help businesses to process unstructured data and organize it according to context. Because data often exists in a range of formats and has to be prepared in a laborious process before anything else can be done with it, having an automated solution is key. The team makes its talking AI smart by fine-tuning how it is fed with the correct information. This gives the AI the capability to answer questions reliably without omitting any important details. As Dr. Jan Bergerhoff explains: “We give the AI access to the companies’ expertise, which allows it to handle tasks in a targeted way. Our service is the first of its kind anywhere in Europe.” To test its platform under real-life conditions and make improvements, Jan Bergerhoff and his team formed ByteIT GmbH, which is already trialling the service with a number of clients.
Open source, privacy and data protection
Adolfo Santamónica adds: “We’re supporters of the open-source philosophy and are keen to improve the quality of open-source technologies together with the community. Among other things, our platform integrates leading open-source ETL tools—i.e. the technologies that let us combine the various data formats in a single system—and develops these further with an eye on tailored use.” Complying with the law and regulations is also important for the founders. “We’re making sure that our solution meets the most stringent European privacy and data protection standards such as GDPR and the AI Act,” Santamónica adds.
Focusing on fairness
The use of AI is increasingly raising ethical questions. With this in mind, the team is focusing squarely on fairness when it comes to programming its AI. Language models inherit the prejudices inherent in their training data when they are being developed, although discrimination can be reduced in a targeted way by supplementing data at the very moment a query is run. Jan Bergerhoff has addressed issues such as gender equality, non-discriminatory AI and non-discriminatory recruitment in his research at the University of Bonn’s Institute for Microeconomics. This expertise puts the team of founders in an ideal position to factor explainable AI and AI security into their work as well.
A successful serial entrepreneur
ByteIT is actually Dr. Jan Bergerhoff’s second start-up project. He previously helped to spin off another start-up, candidate select (CASE), from the University of Bonn, going on to establish it successfully on the market and lead it for several years as its CEO. Before he set up CASE and came up with the new business ideas that would ultimately lead to the Heinzel AI project and ByteIT, he completed a doctorate in Economics from the Bonn Graduate School of Economics.
“Needless to say, it’s always rewarding to collaborate with experienced start-up entrepreneurs, and the team is benefiting enormously from Dr. Bergerhoff’s expertise,” explains Vesna Domuz, a start-up coach from the enaCom Transfer Center at the University of Bonn. “It’s not at all unusual for founders to leave their company once it’s in a good position and turning a profit and go off to develop a new business model.” Explaining what makes serial entrepreneurship unique, she adds: “Thanks to their extensive knowledge of the market, serial entrepreneurs can focus intently on the needs of a particular sector and drive solutions that break new ground. And, more than anything, they can respond quickly to the rapid advances being made in the possibilities afforded by technology—especially AI technology—and the fast pace of market trends.” Domuz helped the team to submit their ultimately successful application to “Start-up Transfer.NRW” in her role as a start-up advisor.
