FinTRACE AI transforms financial analysis with behavioral insights

A new AI system called FinTRACE is set to improve how banks and digital platforms analyse financial behaviour. Developed by researchers at Sberbank’s Practical Artificial Intelligence Center and the Sber AI team, the tool transforms transaction histories into structured knowledge. This allows AI to detect stable patterns and habits before making decisions. FinTRACE combines information retrieval, explainable rules, and large language model reasoning with structured temporal data. It converts raw financial records into a detailed knowledge base that captures individual behaviour over time. This approach helps banks spot potential client issues earlier and offer timely support.

The system is not limited to finance. It can also enhance medical forecasting by analysing data from doctor visits and lab results. Beyond this, FinTRACE works as a universal layer over any event-based data, making it useful for tasks like predicting customer churn or assessing credit risk.

A scientific paper on FinTRACE, led by Maksim Makarenko, has been accepted for presentation at the 49th ACM SIGIR conference. The research highlights how major banks and digital ecosystems can build more flexible, cost-effective, and adaptable AI models using this technology. With FinTRACE, artificial intelligence in financial services is expected to become more accurate, reliable, and transparent. The system’s ability to recognise behaviour patterns could lead to fairer risk assessments and more personalised customer offers. Banks and other industries may soon adopt it to improve decision-making across various applications.