How is AI used in wealth management?

AI in asset management is used mostly for support tasks and rarely for the investment decision itself. In an ESMA survey of summer 2025, 395 firms in EU securities markets reported 833 AI use cases. The largest groups were drafting and summarizing (239 use cases, 29%), internal assistants or copilots (227) and code generation (109). Portfolio risk management (20), portfolio optimization (19) and algorithmic trading (10) were rare, and most of these were still in development or testing.

Funds that advertise AI are a niche. In the first quarter of 2024, 106 EU funds named AI or machine learning in their investment process, with just over EUR 13 billion in assets, about 0.1% of all UCITS assets. ESMA found that these funds "have not delivered significantly higher or lower performance" and that asset managers use generative AI "primarily to support human-driven investment decisions" (ESMA, February 25, 2025).

Wealth managers use AI in client service and in the advice process. ESMA gives examples of AI use cases in wealth management that fall under MiFID II: chatbots that answer client queries, tools that analyze a client's knowledge, experience, financial situation and objectives "in order to provide personalised investment recommendations or manage and rebalance client portfolios", and tools that check whether a portfolio still fits the client's risk tolerance.

Client-facing AI is still the smaller part. In the ESMA survey, 86 use cases served communication with clients and 34 know-your-customer checks, while 87% of all use cases were for internal use only.

Generative AI in investment management is the most common AI type: it underlies 571 of the 833 use cases in the ESMA survey, or 71%.

AI agents in investment management are systems that plan steps and call external tools. They account for 141 use cases, or 17%. ESMA found that 27% of these agentic use cases run with high or medium autonomy, against 19% of all use cases. Across the survey, 77% of use cases run with low or no autonomy, "meaning that human approval is required or the AI system only provides suggestions".

The benefit that firms expect is efficiency, and 70% of the firms in the ESMA survey expect their AI investment to rise between 2025 and 2027. The risks that ESMA names in its statement of May 30, 2024 are over-reliance on AI without human judgment, a lack of transparency and explainability, data security and privacy, and unreliable output. AI tools for text generation "are known to 'hallucinate'", and in advice and portfolio management that "can lead to misleading advice".

The challenges in wealth management are the same as in asset management. Firms in the 2025 survey named data and model weaknesses as a top concern, followed by cybersecurity and third-party dependencies. The CFA Institute published the book AI in Asset Management: Tools, Applications, and Frontiers on November 18, 2025. It teaches practitioners to "balance automation with human oversight".

The EU AI Act does not list investment advice or portfolio management among the high-risk uses in Annex III. In finance, Annex III point 5(b) names AI systems "intended to be used to evaluate the creditworthiness of natural persons or establish their credit score", with an exception for fraud detection, and point 5(c) names risk assessment and pricing in life and health insurance. A bank that scores a client's creditworthiness with AI falls under point 5(b); a portfolio tool is not named in Annex III.

Article 50(1) applies to every client chatbot: people must learn that they are interacting with an AI system unless this is obvious. The general date of application is August 2, 2026. Regulation (EU) 2026/1744, in force since July 27, 2026, moved the start of the Annex III high-risk rules to December 2, 2027.

In Germany and Austria, the AI Act applies directly as an EU regulation, next to the MiFID II conduct rules that BaFin and the FMA supervise. BaFin, which has offices in Bonn and Frankfurt am Main, published non-binding guidance on the ICT risks of AI under DORA on December 18, 2025, addressed to banks under the CRR and insurers under Solvency II. ESMA expects firms that use AI in investment services to comply with MiFID II and to act in the client's best interest.

Switzerland is outside the EU. The AI Act still reaches Swiss providers and deployers when the output of their AI system is used in the EU (Article 2(1)(c)). FINMA published Guidance 08/2024 on governance and risk management for AI on December 18, 2024. It names model risks such as bias and a lack of explainability, data risks and growing third-party dependencies, and asks institutions to contact FINMA in good time before they use AI in critical processes. In a FINMA survey of around 400 institutions, published on April 24, 2025, about 50% used AI or had first applications in development. This page gives no legal advice.