AI in Wealth Management: The Economics of Advice Are Being Rewritten. The Advisor Is Not Disappearing.
CCO Managing Partner
For a few years, the loud version of the story was simple. AI writes the plan. The relationship manager becomes optional. Fees collapse. The house that still employs people starts to look like a museum. I do not think that story survives contact with the place that actually matters: why a client decides to stay.
I sat on the other side of that table long enough. Clients did not pay for a document. They paid for a person who knew the family, could get the right expert into the room and would still pick up the phone when the file was ugly. The pack was the souvenir. The advice was the product. AI changes the cost of the souvenir. It does not delete the product.
What AI Is Actually Rewriting
Three things move at once. First, the cost of preparation falls. Meeting briefs, portfolio narratives, first-pass planning and documentation used to consume a large part of the week. If the house has its data in one place, much of that work can now be done in an afternoon.
Second, the capacity of one advisor rises. Not because the advisor suddenly became wiser, but because the unpaid hours around the conversation shrink. Research from Cerulli and Vista shows wealth firms planning to add more junior advisors, client service associates and senior advisors over the next two years, not fewer. That is not nostalgia. It is a bet that the bottleneck is coverage and relationship quality, not headcount in the abstract.
Third, the fee debate gets louder. If a model can draft the same plan a junior used to build, clients will quite reasonably ask what the fee is actually for. Houses that answer “the pack” will have a problem. Houses that answer “judgment, access and continuity” still have something worth paying for.
BCG describes the fork quite clearly. One path uses AI mainly to disrupt the old cost model. The other uses it to expand advisor capacity while keeping the human core of the relationship intact. For private banking, the second path makes more sense to me, but only if the house is honest about what it is actually selling.
What Does Not Move
Succession. Family conflict. A structure that was legal and messy when it was built. The moment a client asks whether this is the year to sell the company. Those conversations do not fail because the briefing pack was late. They fail when the person in the chair does not know the story or does not own the recommendation.
Capgemini’s 2026 World Wealth Report points in the same direction. The next step in personalisation is not a robot relationship manager. It is an intelligence layer that gives advisors better context, better coordination and more time for the conversation. That is the right instinct. The layer should make the advisor more present, not less necessary.
I have seen the opposite often enough. A beautiful pack. Three versions of the same client sitting in three systems. The relationship manager discovering the family argument in the meeting because it never made it into the CRM. AI will not fix that. It will print it faster.
Where Swiss Private Banks Actually Stand
The Swiss market is a useful reality check. KPMG and the University of St. Gallen found that AI is already widely used across Swiss private banks, but mostly in isolated productivity use cases. The deeper operating model is still missing in most institutions.
The economics tell the same story. Cost savings appear first. Revenue effects barely show up. In the HSG AI Advancement Index, just four of 73 banks reached the Leading category. That is not a reason to wait. It is a warning about what not to confuse with progress.
If you only buy the latest advisor plugin, you have rented a faster junior. You have not changed the economics of advice. The vendor does not have your full client history. It does not know why the last credit was declined. It does not know which specialist the client trusts, which family member refuses to sit in the same room or which introduction actually matters. That knowledge still sits inside the house, often badly distributed.
The Choice Hidden Inside the Economics
Once preparation becomes cheap, the house has to decide what the freed hour is for. Option one: more clients per relationship manager, same conversation, thinner relationship. The spreadsheet likes it. The client notices within a year.
Option two: the same book, better coverage. The advisor finally has time to be in the room for the decisions that used to wait because the file was not ready. That is where I think AI can pay for itself in private banking.
Option three: keep the old process and add a chatbot on top. That is how you pay for the licence and keep the weekend work.
I would take option two. Not because I am sentimental about bankers. Because the part of the job that created loyalty was never the first draft of the quarterly letter.
What I Would Do If I Still Ran a Book
I would put agents on preparation, documentation and the first reconciliation of the client file. I would make every output traceable. And I would insist on one narrative of the client, not three.
I would keep the recommendation, the Source of Wealth conclusion and the family conversation with a named person. If the signature is only formal, you have not rewritten the economics. You have hidden them.
And I would measure the extra hour. If that hour does not show up in meetings that matter, in faster onboarding of complex files or in a client who can reach a human who already knows the story, then the programme is theatre.
Swiss private banks should use their traditional advantages deliberately: trust, continuity and access to a network. Those were always part of the product. They were simply delivered as the habit of good relationship managers rather than as a designed experience. AI can make that habit more systematic. It cannot invent it.
The Sentence I Would Put on the Wall
The pack will be written by a model. The advice still needs an owner. The client will leave the house that forgets the difference.
The economics of advice are being rewritten. The advisor who only produced paper should be nervous. The advisor who carried the relationship just became more expensive to replace, not less.
If you are looking at how AI could change advisor capacity, client coverage and the economics of advice in your own bank, book a 30-minute exploration call with us. We can look at where preparation effort is already falling, which parts of the client relationship should remain clearly owned by people and how to turn freed capacity into better service rather than simply more volume.