Analyze, form a portfolio, forecast profitability. More and more often, investors trust these tasks to artificial intelligence (AI) – to earn money without worrying about the slightest market fluctuations. But how do such models work? What opportunities do they open up for investors and how risky is it to “play” by the rules of a robot? Experts from VTB broker answered these and other questions – the bank became a pioneer in creating an ML product for investors in Russia
In 2020, when AI technologies seemed like a bold hypothesis to brokers, VTB My Investments launched the Artificial Intelligence investment strategy as part of a robot advisor , which is entirely generated by artificial intelligence. Five years later, more than 8,000 clients have trusted this strategy with over 5 billion rubles, and its profitability outpaces the Moscow Exchange index.
The first AI strategy on the market
In Russia, the “Artificial Intelligence” strategy was the first of its kind – with its appearance, investors were able to apply the capabilities of AI in the investment process for the first time.
Over the course of five years, the product has taken root among VTB broker clients and has become one of the most popular investment scenarios: “ The strategy is the second most popular among all VTB My Investments strategies: every fifth client using the robot advisor chooses it. VTB broker clients already have 4.8 billion rubles concentrated in their accounts with this strategy. And since its launch in 2020, the strategy has proven highly effective – clients who followed its recommendations received a total return of more than 44% ,” said Dmitry Melnikov, Head of Investment Management at VTB My Investments .
The expert also shared other data: the strategy confidently outperforms the Moscow Exchange Full Return Index (MCFTR) – since the beginning of 2025, the gap in favor of the AI strategy has amounted to 3.66%. Thus, the ML model (machine learning, ML) outperforms some classical approaches in terms of efficiency – and works even in conditions of high market volatility.
Not a crisis, but a new experience
VTB mathematicians and experts trained the ML model on Russian market data for 2016-2020. This is a unique period in its own way: on the one hand, with a large number of market scenarios – recovery, economic growth, volatility during the COVID-19 pandemic. On the other hand, it is not too “ancient”, which allows making forecasts without relying on outdated patterns.
The model is constantly evolving, but does not need to be “retrained” from scratch. It processes new data and then checks it based on other datasets. As a result, it issues accurate recommendations, does not become outdated, and adapts to real market conditions.
At the same time, the client retains full control over the operations – the robot advisor does not make transactions for the investor, but only generates ready-made recommendations in accordance with the strategy, which the investor can confirm by pressing a button.
But the algorithm does not “learn” from the actions of a specific user. It gives advice based on the market situation, not behavioral patterns. This is essential – so as not to reinforce the mistakes or subjective preferences of the client, but to focus exclusively on data analysis.
A Sustainable Strategy in Two Clicks
The strategy portfolio includes the most liquid shares of 20 Russian companies. Among them are Tatneft, Lukoil, Polyus, Sber and others.
Once activated, the robot advisor analyzes the client’s current portfolio and sends individual recommendations to balance the set of securities in accordance with the chosen strategy. The robot will help throughout the investment process: at least once a month, advise how to adjust the portfolio composition – taking into account market changes.
Risks or the path to high returns?
VTB My Investments experts recommend investing with the “Artificial Intelligence” strategy over a 3-5 year horizon. Therefore, those who are ready to play “long-term” turn to the strategy – including those who are ready for the fact that the value of the portfolio may decrease for some time. As well as young investors who are interested in modern technologies.
In addition, in 2024, the model was recalibrated and independently revised its parameters: it reduced the price analysis window from 8 to 4 months and increased the frequency of rebalancing to respond faster to the changing market. This is further proof that the strategy does not simply follow the algorithm, but adapts to the new reality.
“The potential of the improved strategy has been confirmed by backtests: the updated model on historical data demonstrates an average annual return of about 30% with an expected alpha of about 20% to the benchmark (stock index), which makes it one of the most effective among the robot advisor strategies,” said Dmitry Melnikov .
Investments 2.0: faster work, higher profitability, stronger algorithm
In 2025, the strategy will become even more accessible, said Alexander Kazakov : a mechanism is being prepared to launch that will allow AI to carry out recommendations automatically. In addition, in the third quarter, it is planned to update the product to version 2.0. In it, investors will find:
- Expanded set of instruments. The number of shares for non-qualified investors will increase from 20 to 53. It will be possible to invest in units of the exchange-traded funds “Moscow Exchange Index” (EQMX), “Russian Bonds” (OBLG) and “Gold. Exchange” (GOLD).
- Continuously updating asset weights. Regular updates will help to maximally adapt the strategy to market changes.
- Updated architecture. The model will be transferred to tabular neural networks, which will ensure more accurate and flexible operation of the strategy algorithms.
AI Finds a Place in the Investment Sun
ML models are becoming more modern and efficient, but artificial intelligence is not yet the most widespread tool. This was reported by Alexander Kazakov . According to him, this is due to the fact that the industry is still young, and setting up ML models requires significant financial costs, a large amount of data and competent additional training of models for their tasks. Moreover, the market does not yet use AI models that function completely autonomously. People continue to make key decisions when working with technology.
Artificial intelligence is now actively used in algorithmic trading to analyze historical data, news feeds, and social networks, as well as to predict long-term market trends. In addition, AI is used to analyze crypto markets, including tracking transactions and the activity of major players in this segment. AI plays an important role in risk management and stress testing and, using the example of the Artificial Intelligence strategy from VTB My Investments, helps in forming portfolios.
Another trend noted by the expert is the emergence of AI-based chatbots in investor apps. Robots help clients quickly understand the market situation, explain price movements, and give advice in a convenient dialog format. Such solutions make analytics accessible even to beginners.
In addition, AI is increasingly being used in customer service. In recent years, many banks have managed to implement ML models in the support structure: and now they are transferring answers to typical questions to intelligent assistants.
“Such AI managers are in touch 24/7, answer questions quickly and in the most accessible way. This simplifies work, including with investment products. All this makes investment services more convenient, personal and modern,” explained Alexander Kazakov .
As a result, the ML segment is growing and developing, artificial intelligence is finding new applications among clients. And it is gradually becoming synonymous with “reliability”, “profitability” and “rationality” in investment matters.
