BUT graduates have developed an algorithm that can find start-ups all over the world

Large foreign companies and investment companies who want to find prospective start-ups and innovative technologies are increasingly turning for help to the Czech Leadspicker project. The project employs a smart algorithm based on machine learning and artificial intelligence. It can search the Internet and find just the start-ups the client is looking for. Leadspicker was developed by two BUT graduates - Vlastimil Vodička from the Faculty of Business and Management and Jan Skácel from the Faculty of Information Technology.

The idea to create an automatic tool which can search for start-ups was conceived in Vienna, four years ago. Vlastimil Vodička, an FBM BUT graduate, worked for an Austrian company which helped its clients to discover prospective technologies. "The World Bank tasked us with finding start-ups in the Balkans where there was no database at that time and nobody even knew the word "start-up". Our analysts were manually searching through all databases and commercials registers and I got an idea that it would be possible to automate some of the processes," stated Vodička. He shared his idea with his friend Jan Skácel who previously worked as developer for Seznam.cz and was in the natural language research group at FIT BUT. "I started developing a software which would make searching easier. We found out it was working very well and we could base our business on it," added Skácel, who was motivated by these results to found the Leadspicker project with Vodička.

The algorithm works on the machine learning principle which utilises certain elements of artificial intelligence. It uses selected keywords to search through Internet-based data ranging from websites and databases of companies or start-ups to social networks. Then it calculates how much the found start-up matches the requirements and how would the given technology be useful for the specific client. Leadspicker analysts then re-check the sample and, based on their feedback, the neural network refines its calculation model until it is able to accurately determine which technology companies are most suitable for the client.

"We have managed to minimise the amount of labour required and we are very fast. Our classifier can even find garage-based start-ups founded by some enthusiasts which are not even listed in any database or register yet. We can discover future "Googles" and "Microsofts." And we can do it before anyone else," said Vodička. The quality of the start-up database created by Leadspicker is also based on the fact that it is not static. Start-ups are not only quickly formed but they also quickly cease to exist. For this reason, Leadspicker detects number of visits and activity on websites and social networks and removes start-ups which ceased to exist or changed their focus.

According to the Leadspicker founders, large companies found out it is no longer profitable to invest only into their internal research and development which can take up to several years to come up with its own innovation. "Companies are interested in immediate and effective supply of several dozen new technologies per month. Speed is crucial, they need to innovate their own processes or place new product on the market sooner than their competitors," explained Skácel. That is why companies from many different areas approach Leadspicker - from automotive industry through energy industry, healthcare and financial services to software or consulting companies. Their clients include brands such as PricewaterhouseCoopers, Axel Springer, Vodafone or Swiss Post.

During four years of its existence, Leadspicker grew from its two founding members to a team of fifteen and is still accepting new co-workers. So far, it co-operated with over 150 clients from thirty countries all over the world. "We are growing fast. During the past three years, we have managed to triple our turnover; last year, we have exceeded five hundred thousand dollars," noted Vodička. Both Leadspicker founders also want to use the data collected on start-ups to make investments. "We are in contact with the largest companies in the world and we know what technologies they are looking for. But these companies not always want to buy a start-up in its infancy. That is why we feel it is logical to fill the empty space and make some smart investments at the right time," concluded Vodička who is currently also working on founding of a new investment fund.