While private banks and asset managers are heavily investing in AI for their own processes, something is also changing on the other side of the table: A client who shows up with six pages of AI-generated advice. “It is now essential for managers and banks to make their advisors highly AI-savvy.”
At present, private banks and asset managers are investing heavily in so-called AI toolkits, but there is still too little attention for the role of AI on the advisory side, says Jan Bellens of EY, who as a partner advises financial institutions on growth and transformation, in conversation with Investment Officer. “It is an absolute requirement for managers and banks to make their advisors highly AI-savvy,” he said. “And to gain faster insight into how clients are using it.”
“Suppose a friend has a birthday tomorrow, your digital assistant will send her a bouquet of flowers and prepare a WeChat message to congratulate her”
Jan Bellens, partner at EY
During the years the consultant lived in China, he saw far-reaching examples of the major role AI can play in people’s personal lives. The open source software Open Claw, conceived by an Austrian developer, is being embraced there. People give the system access to their applications, allowing their digital assistant to maintain a personal data cloud. “Suppose a friend has a birthday tomorrow, your digital assistant will send her a bouquet of flowers and prepare a WeChat message to congratulate her.”
The Chinese also use it for their investments, said Bellens, although mainly within the “younger urban professional segment.” He does not expect the fully autonomous AI assistant—which independently finds and executes the best deals—to gain traction in Europe anytime soon. European privacy rules, the AI Act, and financial regulation are slowing that development for now. “But I give this example because it shows where it could go and what will eventually be possible.”
Recommending unsuitable products
The Dutch Authority for the Financial Markets (AFM) takes a similar approach, thinking in scenarios and extremes. During its annual presentation in early April, two AFM experts outlined the AI risks the sector should take into account. They emphasized that these are not yet occurring, but are realistic.
In the case of pre-trade risks, the experts warned about retail investors using GenAI as a substitute for formal investment advice. GenAI may recommend unsuitable strategies or products, potentially involving higher risk than appropriate, and without investor protection.
The AFM study is an exploration, intended to stimulate debate within the sector and among regulators because the topic is still at an early stage. The use of ChatGPT by clients of private banks and asset managers—for example to discuss their portfolio or prepare critical questions—is still rare, according to AFM board member Hanzo van Beusekom.
“You never have full visibility into what a client does,” Bellens responded. “And that is where asset managers and banks need to pay attention. What is happening right now that we do not yet have a clear view of?” EY is researching this, but Bellens notes that developments are moving extremely fast and that surveys do not always reflect what is actually happening in practice.
“The questions that are ‘prompted’ are often good questions”
Frederik Kalff, director at Delen Private Bank Netherlands
A survey among asset managers and banks reveals mixed experiences. At some firms it is “not happening yet,” but at Delen Private Bank, for example, clients regularly arrive with six printed pages filled with questions and conclusions generated by ChatGPT for the investment advisor. Director Frederik Kalff noted: “The main conclusion so far is that it is not very valuable. It often contains inaccuracies, conclusions based on hallucinations. But there are certainly also good elements. The questions that are ‘prompted’ are often good questions.”
Not waiting for an endpoint
Bellens’ advice to banks and asset managers is to ensure that advisors are fully up to date on developments in AI and are able to use it themselves. “And that is a major challenge, just look at the first wave of digitization. Some advisors embraced digital tools, while others stuck to their printed Excel sheets.”
Kalff of Delen added: “I think the most important thing with AI—and I say this internally as well—is that there is no endpoint, and we cannot wait for one. Start with small things. I like to see on the floor here that everyone is open to it.”
When it comes to client advice, private bankers at Delen try to understand which questions clients bring into meetings. They regularly put themselves in the client’s position and “consult” ChatGPT in that role, then analyze the answers and advice the application provides. In Belgium, the private bank has its own innovation hub, where these kinds of applications are tested and further developed, alongside the use of AI to improve internal systems.
“It is of course a very big question, not least because it will put pressure on fees”
Jan Bellens, partner at EY
This aligns with Bellens’ second recommendation: to make advisors aware that clients will use AI tools and that this will affect advisory conversations. The statement that human judgment will always remain important is, in his view, too simplistic. “I think you really have to think more deeply about the value of your advisors, and how you maintain trust in a world where you combine AI tools with human input. And that is of course a very big question, not least because it will put pressure on fees. At the same time, it could also create more work, if you have to address all those client questions.”
Bellens is not surprised by the scenarios market participants are sketching behind the scenes. A situation in which a client places their phone on the table during an advisory meeting and lets their digital assistant listen in and ask questions is, in his view, entirely plausible. “It is impossible to predict exactly where this is going, but it is certainly not standing still.”