MIT Sloan: AI Gives Good Financial Advice If You Ask the Right Questions

A study by MIT Sloan Management Review found that AI chatbots can give financial advice at the level of professional consultants, provided the user asks specific, detailed questions. The work was published in 2026 on the "Ideas Made to Matter" platform and sparked wide discussion in the professional community.

What the study found

MIT Sloan tested language models on typical financial advisory tasks: retirement planning, investment decisions, budgeting and tax questions. The results were unexpected: the AI systems performed significantly better than the authors had assumed.

The main finding: the quality of AI answers varies dramatically with how detailed the query is. A vague question ("how do I save for retirement?") produces general recommendations indistinguishable from any personal-finance article. A detailed query that specifies age, income, current assets and a concrete financial goal is a different conversation entirely. According to MIT Sloan, it is the quality of the question that determines the quality of the advice.

Why did AI turn out better than expected?

Language models showed particular strength in two tasks: analyzing several variables at once and calculating "what if" scenarios. A human advisor often falls back on a standard template for a similar client; the AI works through every condition in the query in sequence, without simplification.

"AI financial advice is surprisingly good, if you ask the right questions," is how the MIT

Sloan Management Review researchers framed their main conclusion.

According to MIT Sloan, AI tools are especially effective in several scenarios:

• Initial evaluation of options before meeting a professional advisor
• Educational tasks, such as understanding how a specific financial instrument works
• Comparing scenarios with numerical parameters (different rates, different time horizons)
• Testing your own hypotheses about investment or retirement planning

The authors also warn that AI advisors are less reliable where precise interpretation of tax law or the legal nuances of financial products in a specific jurisdiction is critical.

How to phrase questions correctly?

The key principle MIT Sloan highlights is to give the model as much context as possible. Age, income, marital status, investment horizon and a concrete goal: the more precise the input data, the more substantive the answer.

The authors also recommend asking about a specific decision rather than a general strategy, asking the model to compare several scenarios using numbers, and stating your country, since legislation differs everywhere. This approach fundamentally changes the nature of the interaction: the model stops generating platitudes and becomes a tool for real analysis that accounts for your circumstances.

What this means

The MIT Sloan study shifts the question from "will AI replace the financial advisor?" to "how can you use it well right now?" The way you phrase a query is the difference between useless advice and analysis comparable to a professional consultation. For most users, the best strategy is to use AI to prepare options and test hypotheses, and to bring in a professional advisor for the final decisions.

Frequently asked questions

Can AI replace a financial advisor?

The MIT Sloan authors do not propose AI as a full replacement. The tool is especially strong at scenario analysis and educational tasks, but it falls short of a professional where precise interpretation of a specific country's legislation or personal accountability for a long-term financial plan is required.

Which financial questions is AI most useful for?

According to MIT Sloan, language models perform best when comparing options: retirement planning, investment scenarios and budgeting assessment. They are weaker in legally complex situations where precise interpretation of the legislation of a specific jurisdiction matters.