istari.ai – the data wizards: Corona brought the first revenue

David Lenz and Jan Kinne generate corporate information in real-time. With their analyses they support companies and European policy makers in identifying the economic consequences.

David Lenz and Jan Kinne generate corporate information in real-time. The Covid-19 pandemic pushed their business model. With their analyses they support European policy makers in identifying the economic consequences.

Jan Kinne, Miriam Krüger, David Lenz and Robert Dehghan (from left to right) form the team of istari.ai

David Lenz (30) and Jan Kinne (31) are Istari.ai. The “webAI” of the startup can generate real-time company information in a new way: it can analyze company networks, evaluate company performance and can be individually customized. The founders got to know each other during a research project and their first joint paper led to the business idea while both were still in the middle of their doctoral theses. A year and a half after the company was founded, the Corona pandemic hit global markets – and with it came Istari’s first revenues. Read here to find out which industries can benefit from web-based innovation indicators and what the next steps are for the start-up.

The way to founding process

Jan Kinne is a researcher at the ZEW (Leibniz Centre for European Economic Research) in Mannheim and has been working with corporate data for many years. When he started his doctoral thesis, he wanted to find out microgeographically where companies are located and how they influence each other. It turned out that traditional innovation indicators, especially patent data or data from company surveys, are not particularly well suited for microgeographic insights. “This gave rise to the idea of using company websites to find out what they do and how innovative they are. I submitted an application for the project TOBI (Text-Based Output Indicators as the basis of a new innovation metric) to the BMBF (Federal Ministry of Education and Research). The idea behind it was to use text data from company websites as a data basis for web-based innovation indicators. To this end, we looked for the Justus Liebig University of Giessen (JLU) as a partner and David was hired for the project,” says Jan. David Lenz is a research associate at the JLU at the Chair of Statistics and Econometrics and studied in Giessen. In 2012 he started working for a data science company in the region and gained a lot of experience in information retrieval from data. He plans to finish his dissertation this year.

The first paper led to their business idea

As part of the research project, the two wrote their first scientific paper and realized “that what we do has a lot of potential, even outside of innovation research. Due to the immense methodological advances that have been made in text analysis in recent years, such as deep learning models, an incredible amount of information can be extracted from websites”. For the founders it was clear that they could reform the way traditional providers of business information work: “There are companies with thousands of employees who do nothing but scan the commercial register and see who’s new. They send out questionnaires and that results in an entry in a database that, in the worst case, is no longer updated. With our solution we can offer something much more modern and exciting. It was clear to us that we wanted to create a company that would do exactly that: generate company information in real time in a new way,” Jan sums up her business idea. In early 2019, when they were both still in the middle of their dissertations, they registered their UG.

The three-year research project enabled David and Jan to take a lot with them for their own company. “That was research front what we did and what we learned there about programming and about web-based business analysis, we can take over exactly at our company. This knowledge is now in our heads – and we use our heads every day,” explains Jan with a smile.

Jan Kinne and David Lenz during research work at Harvard University

 

Corona brought the first sales

When the corona pandemic began, Germany reacted with a lockdown, but no one knew who was affected and to what extent. Politicians had to take immediate, targeted action, but there was practically no data on which they could base decisions. “Our approach was attractive because we were able to provide an overview of the situation very quickly. We continuously scanned German websites twice a week and were able to provide a comprehensive picture of the current situation, which is only possible with our approach. In such a special situation, data from March is already obsolete in April,” Jan describes the situation in April. They noticed a steady increase in the number of reports. Switzerland also showed interest in the data and bought it. With this, the startup received its first revenue. “Receiving the first payment was simply amazing. That motivates you to keep going and gives you the feeling that you are on the right track”, says David. This is how the founders were able to hire their first employee. Robert Dehghan supports the team in development, but is also responsible for PR and social media. In addition, Istari.ai is supported by Miriam Krüger, who represents the young company in Berlin and builds up a network for Istari.

The potential market? Huge!

As with the Corona application, political consulting can develop into a line of business, but also management consultancies or all those who do market analyses could be interested as well. For Jan, there are other potential customers as well: “Of course IHKs would be interesting. There the need for information about all the companies in their region is considerable. But staying up to date costs a lot of capacity”. Other data users, such as investment companies, can also contact the start-up. Rating agencies, for example, calculate credit ratings for companies, like a kind of Schufa for companies, and need a broad database for that. Thanks to the data from Istari.ai, such companies could improve their credit rating models.

 

With their research work, the founding team has already been to the German as well as to the European Statistical Office – so the approach they developed has the potential to become official statistics. They have also presented their research approach to the OECD and the Deutsche Bundesbank, so there is a great deal of interest in their work.

2021: Full speed ahead into the development year

The two founders have many plans for the upcoming year: First, they want to open an office in the Mannheim technology center MAFINEX. In addition, the founders are planning their own platform on which their data can be explored on a dashboard. You search for a company, for example “istari.ai”, and arrive at a profile with a machine-generated summary: Tags such as “machine learning”, “text analysis”, “B2B”, “startup”, “data analysis”, “research” and “artificial intelligence” appear. The profile is complemented by a partner network to see who has mentioned the company or has set hyperlinks to the Istari.ai website. Furthermore, there should be the possibility to filter the tags, for example to see which other companies also deal with topics like AI. The results can also be filtered regionally. The platform should be a free offer, for exports or further analyses a premium account with costs is planned. “What we have developed, we now want to bring to the market as a product. But we will also develop further indicators and open up new data sources, such as social media or news articles,” explains David. “Over the next two years, we’re going at full throttle to see if it works. 2021 will be a decisive year for us, in which we will develop a lot and then dare to launch the product,” Jan summarizes.

Many thanks to Benjamin Stuchly and his team for this article, which was published in the Blog Mittelhessen on October 8, 2020.