Sber is preparing to launch a financial AI assistant based on GigaChat that will be built into the Sberbank Online app. The agent promises clients less routine and help with long-term goals, but a significant portion of analysis and money operations will shift to the algorithm. For shareholders, AI remains a new source of earnings, but the coming years will require large-scale investment.
The bank unveiled its plans at Investor Day on November 12. The new service became part of a global strategy: AI is meant to help the bank earn on personal products and curb expense growth.
The assistant will take on analysis and routine
First Deputy Chairman of the Sber Executive Board Kirill Tsarev said the assistant will appear in the Sberbank Online app. It will carry out everyday financial tasks and help users move toward long-term goals.
More than 12 million clients use the bank's existing AI services every month. Among them are a tool for analyzing expenses and finding savings opportunities, an investment assistant for evaluating a portfolio, and a service for finding expenses for which a tax deduction can be claimed.
The new product will bring the user's income, obligations and financial plans into a single system. Goals and key decisions will remain with the client, while AI will take on data research, option selection and routine operations. In the future, the agent will be able, with the user's consent, to connect third-party services and perform actions approved by the user.
Betting on new revenue and curbing costs
Sberbank expects to gain new sources of revenue with the help of AI. This follows from the bank's statement at the Corporate Information Disclosure Center.
"Sber sees its main opportunities in the further personalization of traditional financial products, as well as in the development of corporate solutions based on generative AI for its partners," the document says.
In other words, the bank wants to earn both from its own clients and from the companies to which it will offer AI solutions. The bank acknowledges that developing the technology will require "substantial investments" in infrastructure and system reliability. Sber expects generative AI to help improve internal efficiency and curb the growth of traditional costs. This will make it possible to keep the cost-to-income ratio (CIR) in the 32–35% range over the next three years: the lower the figure, the less the bank spends on its own upkeep out of every ruble earned.
Sber CEO German Gref said the bank's income from AI over the past three years amounted to 1.5 trillion rubles. At the same time, he warned that the figure will worsen over the next three years due to the need to increase investment in model development.
Where else Sber is betting on AI
A GigaChat-based portfolio breakdown appeared in SberInvestments about five weeks ago: the service shows which securities had the biggest impact on the result and links changes to news. Similar functions are still scattered across individual bank services, and the new agent is meant to bring them together in one place.
Sber Senior Vice President Andrey Belevtsev said two weeks ago that about a thousand company executives are already assigning tasks to "GigaAgent" — from analytics to app development. It is precisely such corporate solutions that the bank calls one of the main sources of future AI revenue.
Cryptalist analysis: Sber's strategy reflects the broader shift of major financial institutions toward agentic AI products, where monetization is built on personalization and corporate solutions rather than transaction fees alone. The stated increase in investment against a target CIR of 32–35% indicates that the market will assess not the speed of the assistant's launch, but the bank's ability to maintain operational efficiency amid a capital-intensive technology cycle.