AI forms the financial preferences of users: which banks of Ukraine fall into the answers

More and more users are turning to chatbots with artificial intelligence not only for information, but also for financial advice. Many people ask ChatGPT, Gemini or Claude where it is more profitable to open a deposit, issue a credit card or choose a bank for doing business. However, a new study has shown that such recommendations always reflect the entire market.

For the first time, the plusone AI agency analyzed the extent to which Ukrainian banks are represented in the answers of popular AI models. It turned out that the vast majority of financial institutions hardly appear in the recommendations, while a few of the biggest players appear in them regularly.

AI Recommendations

The authors of the study studied the responses of ChatGPT, Gemini and Claude to dozens of typical user queries related to banking services. The analysis covered the three most popular categories:

• opening deposits;
• FOP maintenance;
• selection of credit cards.

In all scenarios, the same banks turned out to be the leaders. Most often, artificial intelligence advises PrivatBank, monobank and FUIB. Oschadbank and Raiffeisen Bank are also among the most frequently recommended.

At the same time, the rest of the Ukrainian banks are practically absent in the answers of the models. Even changing the wording of a query or using different modes of operation of chatbots rarely affects the list of recommendations.

Why them

Researchers note that modern language models do not constitute long ratings. As a rule, the user offers a short list of the most relevant options, which is formed on the basis of a combination of many factors.

Positions of banks can be influenced by:

• level of brand recognition;
• scale of financial organization;
• reputation in the market;
• development of digital services;
• quality of internet banking;
• product popularity among customers;
• volume of available information about the bank on the Internet.

That is why the biggest players get a significant advantage over lesser-known market participants.

Web search changes answers, but not always for the better

A separate part of the study was a comparison of the performance of models with and without internet search enabled.

The results showed an interesting pattern: some banks that regularly appeared in the recommendations thanks to internal knowledge of the models lost their positions after connecting the web search. In other cases, the situation was the opposite.

This suggests that modern AI uses several sources of information at once. In addition to the data obtained during training, the relevant content available on the Internet at the time of forming the answer is of great importance.

A new competitive factor

According to plusone AI experts, today it is not enough for banks to develop only their own products or advertising campaigns.

As the popularity of ChatGPT and other intelligent assistants grows, a new task appears – to ensure high AI brand visibility. In other words, the company must be represented in the digital space in such a way that artificial intelligence can easily find relevant, reliable and structured information.

If a bank is practically absent in materials that analyze language patterns, the probability of its appearance in recommendations is significantly reduced, regardless of the quality of the services offered.

The future of banking competition

More and more people are using artificial intelligence as the first source of information when choosing financial products. If models constantly recommend a limited range of banks, users may not even learn about the existence of alternative offers with more favorable terms.

This is why the study shows a new trend: competition is gradually going beyond traditional search and advertising. Now companies have to fight not only for a place in Google results, but also for a presence in the answers of ChatGPT, Gemini, Claude and other AI services that are becoming full-fledged digital consultants.