The conventional view of AI in financial services distribution is that it lifts everyone. Tools get cheaper, capability gets democratised, the small intermediary catches up with the large one. The data is starting to say something different.
AI is not narrowing the gap between sophisticated and unsophisticated intermediaries. It is widening it. And for European asset managers building retail-accessible products on the assumption of a broadly capable distribution layer, that is a strategic problem most are not yet pricing in.
What is actually happening
Across UK advice, and increasingly across European intermediary models, two things are happening at once.
One, firms that have embedded AI well are absorbing more complex products, serving more clients per adviser and pulling distribution share from firms that have not. They are moving up the product complexity curve, not because they want to, but because they can. Their cost-to-serve at the lower end has dropped, and they have the operational headroom to take on instruments that require real suitability work, real ongoing service, real reporting.
Two, firms that have not embedded AI, or have bolted it on without rebuilding the workflow, are doing the opposite. They are offboarding clients, narrowing propositions, and quietly retreating from any product that demands more operational capacity than they have.
Taken together, this is not a story about technology adoption rates. It is a story about distribution share moving in real time from one group of intermediaries to another. The bifurcation separation is already underway.
The evidence is already in the data
Our SOFI by Numbers 2025 work, drawn from advisers using purpose-built AI inside their compliance workflows, found 78% saving three or more hours a week on compliance admin alone. That is over 150 hours per adviser per year. The relevant question is not whether the time saved is real. It is. The relevant question is what the firms saving it are doing with it.
The firms reinvesting that time into client capacity, more complex propositions and product breadth are pulling away. The firms absorbing it as slack are standing still. In a market where Consumer Duty in the UK and the Retail Investment Strategy in continental Europe are both raising the operational bar simultaneously, standing still is going backwards.
Why this is not yet showing up at the asset manager level
Most European asset managers do not see the bifurcation because their distribution data does not show it. AUM by intermediary tells you who held the assets last quarter. It does not tell you which intermediaries are operationally equipped to absorb the next generation of retail-accessible products, and which are about to get squeezed out of the conversation.
The product roadmap looks reasonable on its own terms. Active ETFs, evergreen structures, ELTIF 2.0, semi-liquid private markets vehicles, all built on the assumption that the intermediary layer can carry them. That assumption is being tested, and in some segments of the European distribution map it is already failing.
What a tiered distribution strategy actually looks like
Three things change if you take the bifurcation seriously.
First, you map your distribution exposure to operational capacity, not just to AUM. The question is not which intermediaries hold your assets today. It is which of them will still be commercially capable of carrying complex retail products in three years. That is a different list, and for most asset managers it is a list nobody on the distribution team has been asked to build.
Second, you stop treating product innovation as a substitute for distribution strategy. Building a more complex retail product does not solve the problem of an intermediary layer that cannot operationalise it. In some cases, it makes the problem worse, because the firms most likely to take the product on are the ones least equipped to service it well.
Third, you invest selectively in the distribution partners who are doing the operational work. That means data partnerships, infrastructure investment, and a willingness to underwrite the implementation cost where the strategic logic supports it. The asset managers who do this will end up with a smaller distribution footprint and a higher-quality one. The ones who do not will end up with the opposite.
The board-level question
Most asset management strategy decks still treat distribution capacity as a given. It is not. It is a variable, and it is moving fast, and the direction of travel is not the one the industry has been telling itself.
The honest exam question for any European asset manager building a 2027 retail product roadmap is this: do you know, with any precision, which of your distribution partners will still be able to carry complex products at scale by then? If the answer is no, the product roadmap is the wrong place to start.
AI is not levelling the distribution market. It is sorting it. The sooner that becomes a board-level conversation, the better the position you will be in when the sorting is done.
