Sofascore and Aircash are two Croatian based product companies that started small in Croatia but built a global product used by millions of users worldwide.
So, it was useful to hear experiences from their founders when talking about implementing AI in their business. It was all part of a dynamic panel discussion at the Weekend Media Festival – biggest regional Media and Marketing festival.
Ivan Bešlić, co-founder and CSO of Sofascore opened the floor: ‘’AI became part of our world ten years ago, but at first it was just algorithms. Today, when I buy a washing machine, it is an AI machine. We have a big AI hype. But if you are a product company and buy AI tools without internal knowledge of how it works, it won’t be good for you. Just like a fast car – If you don’t know how to drive it, you will go off the road’’.
Hrvoje Ćosić is CEO at Aircash, and he agrees: ‘’When we started, we didn’t have internal resources. We knew what we wanted to do with AI, but we didn’t have the resources to put it to work. At first, we tried to buy solutions, but now in recent stages we are building internal solutions and hiring engineers.”
AI is an important part of both Sofascore and Aircash.
Sofascore has a 15-person team in the AI department. Artificial intelligence is used as a chat filter, it is also a key factor of personalizing the product for every user. AI is part of internal sport analytics, for extracting valuable information from vast amounts of sport data. But there are also some interesting AI projects in development.
At Aircash artificial intelligence is used for onboarding customers, but most importantly – it is used for risk management, since they are a financial institution.
‘’Now everyone has faith that everything will be OK. But everyone should be a little more pragmatic. For example, when you want to find out when the last ferry is leaving from the place you are at, you will probably go and check at the website of that ferry company, and not ask ChatGPT.’’, Bešlić concluded.
The main message is clear – Without the knowledge of an internal team, you can’t decide what is right or wrong in AI processes.