ST. PETERSBURG. July 5. /TASS/. Sber plans to earn about ten billion rubles ($111.5 mln) on account of the GigaChat, the bank’s own neural network based on generative artificial intelligence, First Deputy CEO of Sberbank Alexander Vedyakhin told TASS on the sidelines of the Financial Congress of the Bank of Russia.

"We already see that sales are growing by 10% when including the GigaChat, for example, into sales processes. This is very sound statistics. Sber itself will earn about ten billion rubles this year on account of generative artificial intelligence, on account of the GigaChat. It is gradually becoming a very good business model," the banker said.

The competition for clients is rather fierce at present, Vedyakhin noted. It is impossible to make a high quality offer to an individual without artificial intelligence. "It means good programs and artificial intelligence are needed, which can provide that. Large language models are the latest word. Development of the large language model is very expensive even for Sber," he added.

Sber in 2023 released its version of the GigaChat multimodal neural network, which can answer questions of users, support dialog, write software code, create texts and images based on descriptions within the single context framework.

The Financial Congress of the Bank of Russia is the Russia’s largest annual international financial forum. TASS is the general information partner of the congress.