The AI Trap Nobody in Private Banking Is Talking About

When the bank's technology gets smarter about your clients, what exactly are you taking with you when you leave?

Let me tell you about a press release I read last week.

On March 26th, Bank of America announced the full-scale rollout of something they are calling AI-Powered Meeting Journey. It is, by their own description, an integrated solution that searches and consolidates client relationship insights and recent activity into ready-to-use prep materials before every client meeting, then generates summaries and actionable next steps afterwards. The stated purpose is to help advisors redirect meaningful time toward strategic planning and deeper client engagement.

On the surface this reads as a productivity story. Advisors get more time. Clients get better service. Everyone wins. But I have been in executive search in private banking for long enough to read what is underneath that kind of language and what I see there is a question that the industry is almost entirely failing to ask.

If the bank's AI now knows your clients better than you do, their history, their preferences, their recent concerns, the last three conversations you had with them, who exactly owns that relationship when you walk out the door?

Portability has always been the private banker's ultimate asset

This is not a new observation, but it bears repeating because it is the foundation of everything. The private banker's value, what makes the best ones virtually impossible to replace and extraordinarily expensive to lose, is not their investment knowledge, their compliance expertise, or even their personality. It is the fact that their clients trust them personally. Not the institution. Them. The phone number the client uses when something goes wrong at 11pm is the banker's mobile. Not the bank's switchboard.

That personal trust is what translates into portability. When a senior relationship manager at a major Swiss private bank decides to move to a competitor, to a boutique, to an EAM, a meaningful portion of their book tends to follow. Industry surveys have consistently shown client follow rates of 40 to 70 percent for senior bankers with established books, depending on the market segment, the nature of the relationships, and the legal constraints in place. It is the single most important number in any private banking hire, and it is the number that hiring committees in Geneva, Zurich, London, and Dubai ask about first.

This is why, in my practice, I spend so much time understanding portability before anything else. Not the AUM headline figure, any banker can claim a number, but the real texture of the relationships underneath it. How long? How personal? Who has the primary contact? Does the client know the bank's brand or the banker's face?

Now, for the first time in decades, that calculus is being quietly disrupted by something that has nothing to do with regulation, nothing to do with non-solicitation clauses, and nothing to do with market cycles. It is being disrupted by artificial intelligence.

The technology is not neutral

Here is the dynamic playing out right now across the major private banking institutions. Banks are deploying AI at an extraordinary pace. According to data from Selby Jennings, around half of Swiss financial institutions are already using AI, with a further quarter planning to adopt it within three years. McKinsey has estimated that generative AI could add between $200 billion and $340 billion annually to the global banking sector, primarily through productivity gains. And AI is expected to boost front-office productivity by 27 to 35 percent by 2026.

The front-office applications being prioritised are not abstract. They are meeting preparation tools, relationship intelligence platforms, next-best-action engines, and CRM systems that build increasingly detailed maps of every client interaction. Every call, every preference, every concern, every product conversation, all of it is being encoded into proprietary platforms that the bank owns and controls. The AI gets better the longer the relationship runs and the more data it processes. It becomes, in effect, an institutional memory for the client relationship.

This is where the asymmetry begins. The banker brings the trust. The bank builds the data. And when the banker leaves, the trust is portable, it travels with the banker in the form of a phone number that clients will still answer. But the data does not. The meeting prep system stays. The relationship intelligence dashboard stays. The complete interaction history that the AI has been trained on stays. The new banker hired to replace the departing RM gets the benefit of that institutional memory from day one.

In other words: the more deeply a bank's AI learns a banker's client relationships, the more the bank, not the banker, becomes the repository of relationship intelligence. And the harder it becomes, psychologically and practically, for a client to fully follow their banker elsewhere.

The banks know this. They just are not saying it.

The public messaging around these tools is carefully constructed around the advisor's benefit. Bank of America's own press release states that the role of the advisor will always remain central to the client relationship and that AI enhances efficiency but cannot replace the valuable judgment, empathy, understanding and personal connection advisors bring to clients. This is almost certainly true in any individual client interaction. A client going through a divorce, planning a business sale, or navigating a generational wealth transfer needs a human. The AI is not going to handle that conversation.

But the structural effect of these tools on the banker's longer-term leverage is a different question entirely, and it is one that the institutions have no commercial incentive to raise. Every layer of AI that encodes a client relationship into a bank's proprietary platform is a layer that makes that relationship marginally more institutional and marginally less personal. Not dramatically, not overnight, but consistently, over months and years.

What concerns me is not that this technology is being deployed. It is that the banking talent market is not yet pricing in its implications.

What this means for hiring, now

According to the KPMG Clarity on Swiss Private Banking study published last year, Swiss private banks now employ more people than at any point in recorded history, over 40,000 full-time equivalents, even as the total number of private banks has roughly halved since 2010. The remaining institutions are larger, better resourced, and increasingly AI-equipped. And the talent profile they are hiring for is shifting.

Selby Jennings noted earlier this year that private banks are now explicitly prioritising relationship managers with what they describe as clean client portability, meaning RMs whose books are not encumbered by EAM conflicts, non-solicitation clauses, or complex institutional entanglements. That emphasis is not new. What is new is that AI integration is beginning to function as a form of soft entanglement that does not show up in any employment contract.

From where I sit running searches across Geneva, Zurich, London, Dubai, Singapore, and Hong Kong, I am already beginning to see the leading indicators of this dynamic. Candidates who have spent five or more years inside heavily digitised private banking environments are sometimes finding that their clients, while personally loyal, have also become comfortable with the bank's digital service layer in ways that create friction when they attempt to move. It is subtle. It does not break a placement. But it is a variable that did not exist in the same form even three years ago.

The institutions least affected by this dynamic, at least for now, are the boutiques, the partnerships, and the EAM ecosystem, precisely because their relationship model is structurally human-centric and their technology stack is lighter. A banker at Lombard Odier or Pictet still operates in an environment where the platform serves the relationship rather than partially substituting for it. That is a meaningful competitive distinction in a world where the AI arms race is accelerating at the bulge bracket.

The question every senior banker should be asking right now

If you are a senior relationship manager at a major private banking institution, and you are watching your bank roll out AI meeting tools, relationship intelligence dashboards, and next-best-action platforms, ask yourself this honestly: are these tools serving my relationship with my clients, or are they gradually embedding that relationship more deeply into the bank's institutional infrastructure?

The answer is probably both. The technology genuinely does make you more effective in the short term. It is not a trap in any conspiratorial sense. Nobody at the bank is sitting in a room designing systems specifically to reduce your mobility. They are designing systems to retain clients and increase productivity. But the side effect of that, the gradual encoding of relationship intelligence into proprietary platforms, is real, and its long-term implications for banker portability have not yet been seriously examined.

The most portable bankers I place are consistently the ones whose client relationships exist most completely in the human domain, in trust built over years of personal interaction, in shared history that the client remembers because the banker made it memorable, in a dynamic where the client's first call is always to the banker's mobile and only secondarily to anything the bank provides. That kind of relationship is still very much achievable. But it requires a conscious effort to maintain it in an environment where every bank is now incentivised to make itself the primary interface.

The technology is not going away. The productivity gains are real. The competitive pressure to adopt these tools is genuine, and any banker who ignores them entirely will find themselves at a disadvantage in client servicing within a few years. But there is a difference between using AI as a tool and allowing it to become the relationship. The best private bankers have always understood that distinction intuitively. In 2026, for the first time, they need to understand it strategically.