Will AI replace financial advisors?

Financial advisors have plenty of competition. For Russian clients, the main alternative is a bank deposit: why invest in anything if you can put your money in a bank and simply collect interest? Real estate takes the second place: you can see and touch an apartment, rent it out, or pass it on to your children. Quite unlike those exchage traded “candy wrappers.” Now a third rival has joined them – artificial intelligence. A client can quite reasonably ask: “Why should I pay an advisor if I can open ChatGPT or Claude, tell it about my money, and have it sort everything out?”

Modern AI can indeed do a great deal: find and analyze data, study financial statements, construct a portfolio, compare investment instruments, and help make sense of taxes and contracts. It can do all of this quickly and at a cost that bears no comparison to the fee charged by a human specialist.

So do we still need financial and investment advisors? Let’s take a closer look.

What people actually bring to financial advisors

I analyzed several hundred cases handled by Russian financial and investment advisors in recent years. This is not a list of services pulled from colleagues’ websites – these are real client situations. If we simplify considerably, they fall into seven groups.

Investment strategy and portfolio

The most obvious part of the work: what to buy and sell, how to assess an existing portfolio, and how to generate current income or build capital over time.

• One woman came in with a portfolio that had lost more than half its value. The money had also been borrowed, and there were almost no liquid funds left – it was too late to build a new strategy; first, the existing situation had to be dealt with.
• Another affluent client was automatically classified as an aggressive investor based on a risk-profile questionnaire. After speaking with him, however, the advisor reached the exact opposite conclusion: the man was older, had already been burned in the market, did not want drawdowns, and periodically withdrew money. On paper, the questionnaire said one thing; in reality, the situation was completely different.
• There was a portfolio of roughly 300 million rubles spread across numerous brokerage accounts: the client would forget to execute a transaction or simply be away on vacation. At the same time, he would periodically act on investment ideas received “elsewhere.”
• In another case, a person estimated his portfolio at around $5 million. After everything was recalculated, it turned out to be $15 million across 84 positions – a good reason to consult an advisor.
• There is also the opposite kind of request: “I don’t want to figure all this out – just tell me what to buy.”

Banks, brokers, transfers, and payments

For Russian clients, this has become a major area of work in recent years. Where can an account be opened? Can money or securities be transferred there? Why did a bank or broker process a transaction yesterday but suddenly refuse it today? What should be done with assets that cannot be sold?

• The owner of real estate in the UAE was receiving rental income in dollars. It turned out that transferring the money the way he had before was no longer possible. Consultants went through several options. Unfortunately, none of them worked at the time.
• Another investor lived in the Czech Republic and held about $100,000 in shares of American companies; the securities were blocked. The issue was no longer the quality of the investments: it was necessary to understand the custody chain, the relevant details, the owner’s status, and the capabilities of specific intermediaries.
• Other cases involved transfers of large sums from Germany to Russia, inter-depository transfers, an inability to sell units of foreign funds, SWIFT restrictions, and payouts under foreign insurance contracts.

The main difficulty here is that information becomes outdated very quickly. What worked a month ago may no longer work today, while the bank’s website may still carry the old description of its services.

Taxes, reporting, and regulation

Investment assets rarely exist independently of their owner’s tax status.

• In one case, the Russian tax authorities saw signs of a controlled foreign company in a foreign investment-linked life insurance policy. The client, of course, did not actually own a foreign company. It was necessary to determine what exactly the tax inspectorate had identified, which documents to provide, and how to explain the nature of the contract correctly to the Federal tax service.
• In another case, a person lived and paid taxes in Germany, wanted to invest in American ETFs through a U.S. broker, and was concerned about a total tax burden of around 27%. This required looking at German and U.S. tax rules together, filing applications with the relevant authorities, and analyzing the funds’ jurisdictions and possible alternatives.
• The cases reviewed also included legal entities that wanted to invest temporarily idle cash through a brokerage account. Foreign accounts that had to be reported properly. Losses, coupons, dividends, changes in tax residency, and the creation of a personal fund for assets worth more than half a billion rubles.

Family, inheritance, and capital protection

Money does not exist only inside a brokerage account. People marry, divorce, have children, relocate, and die.

• One of the most illustrative cases involved a woman preparing for divorce and trying to determine how much money was held in the family brokerage account. Her spouse handled all the investments, and she effectively had no access to the account; the first step was simply to determine which broker held the assets.
• In another case, heirs were turned away by several notaries when trying to divide Russian securities between a wife and children. Consultants had to determine who should allocate the marital share, what the inheritance certificates should look like, and what the broker or registrar should do next.

Insurance, large insurance payouts, transferring capital to children, and inheriting foreign accounts form another major area. Here, the advisor quickly finds himself at the intersection of finance, law, and family relationships.

Business, real estate, and debt

For a Russian entrepreneur, the financial portfolio may represent only a small part of total capital. Traditionally, most of the money is held in the business and in real estate.

• A company with revenue of around 7 billion rubles could not obtain a bank guarantee of the required size – banks kept refusing. The advisor looked for alternative banks, structures, and relevant contacts.
• Another company wanted to invest idle funds through a brokerage account. It might seem like a straightforward investment task, but it required taking into account the business’s liquidity needs, the timing of future payments, taxes, and restrictions applicable to a legal entity.
• Other cases involve the sale of property abroad and the need to receive the proceeds safely, mortgages, loans, business currency risk, and choosing a country in which to register a company. Sometimes the situation looks like this: a person earns very well for decades, business turnover is impressive, and personal capital is almost nonexistent.

Fixing mistakes already made

A significant share of clients come not to build a system from scratch, but to deal with the consequences of decisions already made.

Inefficient discretionary management that is expensive to exit. A complex structured product that turned out to work very differently from what the client thought when buying it. Frozen foreign assets. An excessively risky portfolio. Loans taken out for investing. Fraud.

In such situations, the question “what should I buy?” may not come up at all. First, one has to understand what the person owns, what obligations already exist, and how much it will cost to put things in order.

And finally, the human factor

Even a good recommendation still has to be implemented.

• An advisor recommended buying a security at, say, 210 rubles – the client executed the trade later at 222. A few months later, the same thing happened again: the reference price was around 37, while the actual purchase was made at roughly 47.
• Another client with a portfolio of around 60 million rubles did not want to place orders himself and asked for the process to be organized for him. That raised questions about powers of attorney, access to the trading system, and transaction confirmation.

Some people stop responding to the advisor despite having an active contract. Others experience changes in their life circumstances. Some intentionally fill out the risk-profile questionnaire in a way that gives them access to potentially higher-return instruments, only to discover at the first drawdown that they are completely unprepared for the risk.

What can AI replace here?

Unfortunately for most consultants, AI can already do a great deal from a technical standpoint: parse a brokerage report, compare a portfolio’s structure with the client’s goals, build a financial plan, review a multi-page contract, find relevant tax rules, cross-check banks’ requirements, analyze a company’s financial statements, and collect data from hundreds of sources. All of this can already be entrusted to artificial intelligence with relative ease. And this is only the beginning.

The value of simply saying “I know a lot about financial markets” is now zero.

If a consultant’s job comes down to explaining what an ETF is, talking about diversification, or putting together a standard investment portfolio, the question “what exactly am I paying a person for?” becomes entirely reasonable.

Moreover, one can simply “feed” the machine multi-page company reports, market data, news, earnings-call transcripts, presentations, and industry statistics – and the program will process in minutes what might take a person days.

However, the ability to quickly find and process large amounts of data does not yet mean the ability to earn consistently.

• An NBER study showed that funds using AI initially did outperform comparable funds that did not. But over time, the advantage narrowed. This was true even for those that adopted the technology earlier than others. The authors’ conclusion is nuanced: AI can generate excess returns, but the effect is not unlimited.
• The picture is even less clear with fully autonomous trading systems. In a review of 77 studies of trading agents based on large language models, only 19 met the minimum criteria for a proper test. Only one accounted for transaction costs, another had a survivorship-bias problem in its sample, and no fully reproducible studies were found at all. Besides, strong historical results do not guarantee future profits.
• The Alpha Arena experiment is telling: several modern models were each given $10,000 and allowed to trade cryptocurrency futures independently using the same initial data. The results diverged sharply: one model more than doubled its capital, while another lost more than 60%. This was a short test in a narrow market segment, and one should not draw broad conclusions from it. But a reliable “money” button powered by a language model has clearly not emerged either.

Perhaps in ten years the picture will be different. But today, we do not have convincing evidence that would allow us to say with confidence: “just connect a neural network to a brokerage account, and it will beat the market consistently for years after fees and with risk taken into account.”

What is happening to the financial advisory market

If AI were already making advisors unnecessary, we should be starting to see it in countries with mature financial advisory industries. But in the United States, the market is growing instead.

According to the Investment adviser association, advisors registered with the SEC served 68.4 million clients in 2024. In 2025, that number grew by 7.7% to 73.7 million, while the number of registered advisors reached 16,544. These are broad regulatory statistics, not a count of personal advisory relationships, but they certainly do not suggest an industry collapse following the widespread adoption of AI.

Surveys point in the same direction. According to Cerulli, even among those who consider digital tools important, only 36% are ready to switch entirely to online advice, while 46% want to retain human involvement. Among investors aged 50+, the share favoring a fully digital model is predictably lower.

A global EY study found an interesting split: 60% of affluent investors expect their financial firms to use AI, but only 28% trust AI as much as a human. In other words, clients want advisors to use new technologies – but they do not want those technologies to replace the advisor entirely.

Paid investment research also remains in demand. People continue to subscribe to analysts, newsletters, and research services, even though they could technically upload financial statements to an AI tool themselves and ask it to analyze them.

For now, the market seems to be moving more toward the “advisor with AI” model than the “AI instead of advisor” model.

Why is that?

A tool is not a guarantee of results

Take a hammer, for example – an excellent tool. You can use it to drive nails or try to tighten a screw; you can break a window or hurt someone. The hammer itself does not become good or bad because of that: the result depends on the task and on who is holding it.

Another example is running. People have known about the benefits of physical activity for thousands of years, while recreational running became a mass phenomenon beginning in the 1960s. Almost anyone can put on running shoes and go for a run: it is inexpensive, the benefits are well established, and doctors and sports organizations recommend it.

So, did everyone start exercising? Not quite. According to data from a global WHO study published in The Lancet global health, almost one-third of the world’s adult population – around 1.8 billion people – did not meet even the minimum recommendations for physical activity in 2022. And over the previous twelve years, that share increased. Knowing that exercise is beneficial and having access to it does not mean that a person will exercise regularly.

It is unlikely that managing personal finances will be any different.

Having ChatGPT or Claude does not turn a person into the chief financial officer of their family. Managing money properly requires more than simply asking a neural network a question; one has to go through an entire process: gather data on one’s assets, formulate goals clearly, account for specific circumstances, ask the right questions and verify the answers, make a decision, implement it – and, most importantly, repeat this cycle for years.

Here are several typical situations in which AI alone is clearly not enough:

• A woman preparing for divorce does not know where the family brokerage account is held. She would be happy to ask AI to divide the portfolio. But first, the portfolio has to be located.
• A person thinks of a complex financial product as a “deposit with a higher return” – why would he ask AI to analyze the risks if, in his mind, it is “just a deposit” and “everything is insured by the Deposit insurance agency”?
• An investor received a recommendation to buy a security at 37, then remembered it only when the price reached 47: the problem here is not a lack of information, but a lack of time and discipline.
• The owner of a large portfolio misses trades, acts on someone else’s “trading ideas,” and does not keep track of his accounts – one more piece of advice from AI is unlikely to help much.

Statistics on ChatGPT usage support this point. According to a 2025 OpenAI study, three quarters of user requests fell into just three categories: practical advice, information seeking, and writing. Around 49% of messages were requests for an answer or advice, and 40% were requests to perform tasks. More complex use cases, such as in-depth analysis or systematic planning, remain relatively uncommon. In 2026, the share of work-related tasks is growing, but outside their core professional activity, people still use AI most often to find explanations.

Giving someone a powerful tool does not mean they will regularly use all of its capabilities. The same has been true of the internet, spreadsheets, low-cost index funds, and online brokers. The technical barrier keeps falling; the human one remains.