When Clients Have AI, Wealth Management Has to Change
From AI-powered second opinions to more demanding personalisation, we look at how technology is changing the relationship between wealthy clients and their advisers.
A client receives a proposal to increase her private-market allocation. Before speaking to her adviser, she asks an AI assistant to explain the strategy, test the assumptions behind it and compare it with keeping the money liquid. Then she adds her existing fund commitments and asks what happens if distributions remain weak for another two years.
By the time the meeting starts, she does not need a lesson in private equity. She wants to know why another commitment makes sense now. That small change in behaviour matters more than it first appears. Wealth managers have spent years discussing how artificial intelligence will change their own businesses. Their clients have already started using it.
An EY survey of more than 18’000 consumers across 23 countries found that 49 percent had used AI to support savings or investment decisions in the previous six months. Twenty-one percent had used AI agents for financial-product recommendations, while 14 percent had already allowed AI to select financial-services providers on their behalf.
A retail investor asking ChatGPT about an ETF is not the same as a family managing several hundred million francs. Even so, one habit is spreading across both groups. Financial information no longer has to be received passively. It can be questioned, rephrased, compared and challenged in seconds. For private banks and wealth managers, that changes where the value in the relationship sits.
The second opinion has become much cheaper
A recommendation from a private bank once took effort to test properly. A sophisticated client could read the documentation, call another adviser, search for competing products or ask someone in their network. Most people still needed time and access to expertise before they could challenge the investment case in any depth.
A client can ask what assumptions sit behind a structured product, what the downside looks like if volatility moves sharply or how a recommended fund compares with lower-cost alternatives. She can upload public documents, ask where the argument looks weakest and arrive at the meeting with a much sharper set of questions.
The answer can be not perfect. Generative AI can misread documents, miss context and sound certain when it is wrong. A client does not need it to replace an adviser, however. She only needs it to become better at testing what the adviser says.
That is already happening alongside another change in wealth management. EY’s 2026 wealth research found that wealthy clients manage around 29 percent of their assets themselves and use 2.3 wealth managers on average. Forty-five percent expect to move between a quarter and half of their assets.
Recommendations now travel beyond the meeting room. A generic market view is easier to compare. A complicated product is easier to ask questions about. An explanation filled with jargon is easier to translate into normal language. The adviser remains important. Simply knowing more than the client is becoming a weaker proposition.
In Switzerland, the harder problem is context
Swiss wealth management makes that shift especially interesting because the average client is unlikely to have a neat financial life contained within one institution.
Banks in Switzerland managed CHF 10’119.5 billion in the first months of 2026, the first time the figure had crossed CHF 10’000 billion. Foreign-domiciled customers accounted for 45.9 percent of assets at the end of 2025, while Switzerland remained one of the world’s largest centres for cross-border private wealth, with around CHF 2’945.7 billion in cross-border private-client assets.
Take a client who wants to know whether she has too much exposure to US technology. One bank can calculate the technology exposure inside the portfolio it manages and produce a perfectly accurate number. It may not see the technology ETF held elsewhere, venture-capital commitments, company shares inside a pension portfolio or the operating business that still accounts for most of the family’s wealth.
A more capable AI model does not solve that. Give it incomplete information and it produces a more sophisticated analysis of the same incomplete picture.
That is where wealth management starts to separate from financial search. Someone who can understand a hedge-fund structure in seconds has less reason to pay a premium for the explanation alone. She has much more reason to value the adviser who notices that the proposed investment becomes uncomfortable once fund commitments held at another bank are included.
For Swiss private banking, consolidated wealth data is no longer only about prettier reporting. It decides how much of the client’s financial life can actually enter the advice.
Personalisation is becoming easier to question
Private banking has used the word personalised for decades. Clients are becoming better equipped to work out how personalised the service really is.
Capgemini’s 2026 World Wealth Report found that only 17 percent of high-net-worth individuals considered their advisory experience seamless and personalised. Forty-two percent said they had to repeat their goals and preferences to the same firm. At the same time, 97 percent of wealth-management firms still primarily segmented clients according to wealth bands.
Assets under management tell a bank how valuable a relationship is commercially. They reveal much less about the person behind the account.
Two clients with CHF 10 million can require completely different advice. One has just sold a company and is sitting on cash. Another still owns an operating business, has property in several countries, private-market commitments coming due and children living abroad.
Their balance sheets may look similar on the first page of a client file but their decisions do not. AI raises expectations because it makes more individual context usable. Previous interactions, portfolio information, preferences, documents and recent changes can all feed into the analysis much faster than they could through a purely manual process.
Clients are not necessarily asking for an algorithm to take over. McKinsey’s 2026 survey of around 5’500 wealth-management clients across Western Europe found that 26 percent of high-net-worth individuals were comfortable using AI, compared with 13 percent of affluent clients. Only 36 percent of HNWIs were comfortable with a fully digital investment platform without assistance. Among private clients, the figure was 22 percent.
When McKinsey asked where AI could improve investment advice, clients placed timeliness first, followed by explainability and personalisation. The interesting part is not that clients suddenly want less human contact. They want less generic human contact.
Slow service is becoming harder to excuse
There are good reasons for some wealth-management decisions to take time. A succession question spanning several jurisdictions should not be answered in five minutes. Nor should a decision involving tax, an operating company, family members with competing priorities and a portfolio spread across several banks.
Waiting three days because somebody needs to find the latest report is different. So is repeating information the bank already holds. Or receiving a portfolio analysis that ignores assets held elsewhere. Or sitting through twenty minutes of broad market commentary before reaching the decision the client actually came to discuss.
Clients now experience a faster standard outside wealth management, and that makes old service friction more visible. Swiss firms know it. PwC Switzerland’s 2026 industry survey found that 52 percent of respondents used AI technology daily, while another 42 percent were exploring potential applications. Firms reported using it across onboarding, personalised recommendations, suitability checks, reconciliation and reporting.
The client increasingly has AI. The adviser increasingly has AI. Once both sides can summarise documents, retrieve information and generate a first-pass analysis quickly, speed alone stops being much of an advantage.
Good advice starts after the obvious answer
A family deciding whether to sell a company rarely suffers from a lack of information. The same is true for an entrepreneur deciding how much liquidity to retain after a transaction, parents considering how to transfer wealth to their children or an investor choosing between another private-market commitment and keeping more cash available.
AI can model scenarios. It can retrieve documents, expose inconsistencies and produce questions the client had not thought to ask. What it cannot do is remove the trade-offs.
Cross-border wealth makes that clear very quickly. A client may live in Germany, bank in Switzerland, own property in France, hold a company elsewhere and have adult children in the UK or US. A financially attractive answer in one part of that structure may create a tax, succession, liquidity or family problem somewhere else.
That is where the adviser earns the relationship. Less value now sits in explaining what an asset class is or repeating information the client can retrieve independently. More sits in knowing which pieces of information belong together, spotting the conflict that a narrow analysis misses and making a recommendation that still works once the client’s wider circumstances are included.
Accountability matters too. When AI contributes to a recommendation, a client should be able to ask which data went into it, which assumptions mattered, what the system could not see and who checked the conclusion. An answer can sound convincing long before it deserves to be acted on.
So what should clients expect from their wealth manager now?
Not necessarily a chatbot. A better starting point is simpler. Technology should remove the work for which clients have little reason to pay a private-banking premium.
Routine information should be easy to retrieve. An adviser should not repeatedly ask for details the institution already has. Portfolio advice should take assets held elsewhere into account where those assets affect the decision. Recommendations should stand up to comparison. Clients should understand when AI has contributed to an analysis and who remains responsible for the final advice. Personalisation also needs to mean more than a wealth band and an old risk questionnaire.
For a Swiss wealth manager, the strongest proposition increasingly begins with a sufficiently complete view of the client to understand what one new decision changes elsewhere. This is important in a country whose private-banking industry manages enormous pools of international wealth precisely because its clients often bring complicated financial lives with them. AI removes some of the easier work surrounding those relationships. It gives clients faster access to information, lets them test recommendations before accepting them and makes generic advice easier to spot. What remains valuable is harder to automate: seeing the whole financial picture, understanding where the conflicts sit and being prepared to make a judgement when several perfectly reasonable answers point in different directions.
