
Most wealth firms now use some form of artificial intelligence, but far fewer use it well. A September 2026 study by Cerulli Associates and Vista Equity Partners placed just 12% of the 68 advisory firms and multi-family offices it surveyed in its top "Leaders" tier (1).
That gap is one of the most useful ways to understand AI in wealth management in 2026. We use the term New Intelligence for these tools, and this piece looks at what firms are doing with them now rather than what they might do one day. We cover how wealth managers use intelligence today, what separates the firms getting results, where agents fit, how investors use it themselves, and where regulation stands in the EU, UK and UAE.
Mostly behind the scenes. The most common uses are preparing for client meetings, taking notes, drafting follow-up communications and clearing administrative work, so that advisers spend more of their time with clients (1).
The early results are practical. In the same research, 64% of firms said AI had reduced manual and administrative work, and 46% said it had improved the quality of their client communication (2).
In the UAE, the emphasis reflects the client base. Advisory firms in the DIFC describe using machine learning and analytics to process research, run scenario analysis and spot risk exposures across multi-currency, multi-jurisdiction portfolios, including liquidity planning and cross-border structures (3). Adoption is rising quickly there too. The Dubai Financial Services Authority's 2025 survey found 52% of authorised firms using AI, up from 33% a year earlier (4).
For a client, most of this is invisible. What you notice is a better-prepared meeting, a faster answer and a clearer picture of a complicated portfolio.
Discipline matters more than budget. Cerulli grouped firms into three tiers: Leaders (12%), Scaling (38%) and Exploring (around half) (1,2). The Leaders stood apart through their operating habits: a named owner for AI, formal training and proper measurement of results.
The differences are wide. 75% of Leading firms run formal AI training, compared with 21% of Exploring firms, and half of Leaders have measured cost savings, compared with 9% of Explorers. Investment is growing across the board, with AI's share of technology budgets expected to rise from 8% to 15% (1).
Regulators see the same gap. The DFSA found that 21% of DIFC firms lacked clear accountability or oversight for AI, even where it was critical to their operations (4). For anyone choosing a wealth manager, how a firm governs its tools says more than which tools it has.
An assistant answers a question, while an agent completes a sequence of steps. An agentic tool might search a firm's internal systems, extract the relevant details and prepare the next action, with a person checking the result before anything happens.
Wealth firms are starting with contained, repetitive work. At an industry panel in October 2025, firms described agents that speed up account opening, and others that pull key contract terms from internal document stores into a spreadsheet for a human to review before anything is recorded (5). Panellists expected 2026 to be the year agents spread more widely.
The most important step is preparation. Wealthspire, a US advisory firm, spent a year cleaning its data and documenting more than 40 operational processes before deploying agents (5). Intelligence is only as reliable as the records it works from, which is why the firms moving fastest often began with their filing.
The current evidence points the other way. Firms in the Cerulli study plan to add client-facing staff over the next two years: 73% expect to hire junior advisers, 67% client service associates and 56% senior advisers (1,2).
Firms are using intelligence to take on preparation and paperwork, and putting people where judgement and relationships matter most. As mentioned before, for clients, that should mean more adviser time. Some operational roles will change, and industry leaders have been open about that (5).
Clients are not waiting for their wealth managers. In an April 2026 survey of 2,100 people across 19 countries, run by the AI provider BridgeWise, 78% said they already use AI tools for investment-related questions, and 46% do so always or often. The Middle East ranked first in the survey's index of optimism about AI in wealth (6).
Habits vary. 58% of 18 to 35-year-olds were frequent users, compared with 27% of those over 50, and women made up 62% of respondents who had never used AI for investment information. Whatever the starting point, the questions that follow tend to be familiar: what is happening, what it means for my family, and what to look at next. A good wealth manager should meet clients wherever they begin, whether that is a chat window or a phone call.
Rules are arriving in stages, and much of the detail is still to come.
The EU's Digital Omnibus, which entered into force on 27 July 2026, postponed the AI Act's stricter rules for "high-risk" systems. For stand-alone systems they now apply from 2 December 2027 (7). Credit scoring is among the named examples. The Act's transparency obligations were not delayed and have applied since 2 August 2026, including the duty to tell people when they are interacting with an AI system (8).
The UK has no AI-specific rules for financial services. The FCA's position is that its existing rulebook, including the Consumer Duty, already covers firms' use of AI. In January 2026, the Treasury Committee called on the FCA to publish comprehensive, practical guidance by the end of 2026 (9).
In the DIFC, the DFSA says it aims to balance innovation with integrity, keeping oversight proportionate while it develops further guidance. Firms there have asked for clearer, harmonised expectations across the UAE (4).
The common thread across all three is accountability: whichever tools a firm uses, it remains responsible for the outcome.
We are actively exploring how intelligence can support the way we manage wealth. The direction we find most valuable looks across a client's whole situation. It would notice the life events that change a family's finances, such as a marriage, a new child, an inheritance or a move abroad, and bring them to the client's and adviser's attention early.
We do not offer this today. The regulatory framework is still taking shape, and we will offer intelligence-led services only within a regulated, auditable structure. They would sit alongside human advice, with an adviser available whenever a client wants one or the decision calls for it.
Can AI give me investment advice?
Within a regulated wealth manager, intelligence tools support the adviser, and the firm stays responsible for any advice given. In most major financial centres, investment advice requires authorisation whatever tools are used. General-purpose AI tools can explain concepts, but they are not regulated and do not know your full circumstances.
Does my wealth manager have to tell me if they use AI?
It depends on where the firm is regulated and how the tool is used. In the EU, people must be told when they are interacting directly with an AI system, such as a chatbot (8). The UK has no AI-specific disclosure rule, though the Consumer Duty may push firms towards transparency (9). The simplest approach is to ask.
Is it safe for my wealth manager to use AI with my financial data?
It can be, if the firm governs it well. Ask where your data is stored, whether it is used to train outside models, and who reviews what the tools produce.
Should I use ChatGPT for investment research?
Many investors already do (6), and general AI tools are good at explaining terms and summarising information. They can be out of date, can state errors confidently, and know nothing about your tax position, family structure or goals. They are most useful as a starting point for questions to take to an adviser.
Most wealth firms now work with some form of New Intelligence. The ones getting real value know who owns their tools, train their people and measure the results. If you want to judge how a firm uses intelligence, ask how it governs it. The answer will tell you a good deal about how it looks after everything else.
Sources