Trending
A friend sent me the new BCG white paper on how AI is rewiring wealth management, and I want to be fair to it. This is a good paper, and the direction is right. AI is going to reshape the delivery economics of advice, and any firm telling itself otherwise is going to get run over. I will go further than BCG did on that point. Financial planning, portfolio construction, servicing, compliance, that work is getting commoditized faster than most owners want to admit, and pretending your version of it is special is not a strategy…it’s a fantasy.
So this is not me crapping on AI. I use it every day. This is me pointing out that the paper way oversells the advantage and stays quiet about the cost.
Here is the part BCG buries in their own artwork. Look at Exhibit 2: they sort the entire value chain into what AI can drive and what stays human, and the human column, the part they admit AI barely touches, is referrals, coaching, multi-generational strategy, and the trusted advisor relationship.
That is their conclusion, not mine. The durable moat is the human trust layer. They spend most of the paper selling the machine and then quietly concede that the machine cannot hold the one thing that actually keeps clients. That is exactly where CIBYC lives, so when people ask me whether AI is coming for the referral, I tell them they have the direction backwards.
The temptation is to say AI cannot do judgment, that it is fake judgment and that it misses the human so it will never touch the real work. I understand the instinct, and I half agree with it, but be careful, because that argument will get you tuned out by any client who is already using an AI tax tool and liking it.
Here is reality: AI does not need to replicate human judgment to commoditize it.
AI only needs to be good enough at a lower price. The steam loom never replicated the weaver. It just made the weaver’s craft irrelevant for most cloth. A lot of what advisors have been calling judgment for years is pattern matching wearing a suit, and AI is very good at pattern matching. That layer is going to get exposed and repriced, and honestly it should. It is the same thing I have been saying about fee opacity. AI reveals what was never as bespoke as it was billed.
The great news is that there is a kind of judgment AI structurally cannot enter, and it is worth naming precisely because the vague version is weak. It is relational judgment, the judgment where the relationship itself is an input.
-
Reading the fear behind the question.
-
Knowing the family dynamic that is not in any document.
-
Weighing one specific human’s contradictions and history and values and being someone that person can look in the eye when a call goes wrong.
AI cannot do that, and not because it is not smart enough. It cannot do it because it is not a party to the relationship. That is a mechanism, not a mood.
Here is where the newest sales research backs the whole thing up, and it comes from a corner you would not expect. Brent Adamson built his career telling sellers to challenge their customers. He wrote The Challenger Sale. His new work with Karl Schmidt, The Framemaking Sale, lands somewhere that should make every advisor sit up.
After all the modeling, the single biggest driver of a high-quality, low regret decision was not the customer’s confidence in the seller. It was the customer’s confidence in themselves. Adamson says it swamped everything else in the data, and it beat the insight effect that made his own name.
Sit with what that means for our world. The client who will not move is rarely doubting you. They are doubting themselves. Am I asking the right questions? Will I screw up my family’s future? That hesitation is self-doubt, not a referral problem you can close with one more credential on the wall.
Now put AI on top of that. The four forces Adamson blames for wrecking buyer confidence are more information, more options, more contradiction, and more risk of getting it wrong. AI pours gasoline on all four. It can generate infinite analysis, infinite scenarios, and infinite second opinions, and every one of them raises the noise the client has to sort through alone. More information does not create decision confidence. Past a point, it erodes it. By Adamson’s numbers, something close to half of stalled deals never go to a competitor at all. They go nowhere, killed by the buyer’s own indecision.
That is the piece BCG’s machine cannot touch. A referral is not a transfer of information. It is a transfer of confidence.
When someone a client already trusts hands them to you, they are not forwarding a data sheet. They are lending you their own certainty, so the client can borrow the confidence to decide. That is the whole idea, and it is why I called it Can I Borrow Your Car in the first place.
The scarce asset in an age of infinite information is not another analysis. It is the trust that lets a person actually act on one.
Which is why I will not walk this line back. Referrals have survived every information technology we have ever thrown at them:
-
Telephone
-
Email
-
The web
-
Social
-
Robo advice
Each wave automated some layer of information transfer, and not one of them displaced the warm introduction for complex, high-trust, high-consequence decisions. For complex work, they did not just survive; they came out stronger, because every wave raised the noise and made the trusted introduction scarcer.
AI is a bigger disruption than any of those, and I am not going to pretend it is not. But it disrupts in the same direction. It commoditizes the advice around the relationship and leaves trust as the scarce input. Same mechanism as the last five disruptions, bigger tailwind this time.
So here is where I stand, and it is not what BCG is selling. The firms that win are not AI-first…they are referral-first, supported by AI. They are AI in the back office and humans at the front door.
Here is the cost that the paper skips entirely:
The more your differentiation depends on an external tech stack, the more you inherit that stack’s fragility.
Lock-in pricing and increased pricing, access changes, and the quiet punchline that everyone buying the same agents erases the very differentiation those agents were supposed to create. Resilience is not about avoiding the tools. It is about refusing to outsource the thing that makes you distinct to something that can break, reprice, or get sold to your competitor next quarter.
Use the machine for everything the machine is good at and be VERY careful how much you rely upon it. Stay human where being human is the whole point.
Be real. Be human. Be authentic. Be alive.
Related: AI Is Making Human Relationships More Valuable Than Ever