Everyone is talking about AI. Few people talk about what sits underneath it.
And that's where the real game is won.
The best AI in the world is worthless if it feeds on bad data
It's a line we repeat often at Abbove, because it holds true every time we look under the hood of an institution.
In a typical private bank, wealth data is everywhere and nowhere at once. It lives in the CRM. In the core banking system. In the portfolio management tool. In spreadsheets on the advisor's hard drive. And in their head, too.
Fragmented. Duplicated. Sometimes wrong. And almost never seen in the context of a whole family.
Putting AI on top of that is like fitting a race engine to a broken chassis. It makes noise. It doesn't move.
Four questions to ask before launching an AI project
Before starting any AI project, we look at four things.
One, the single source. Is there one place where all of a client's wealth data comes together, or do you have to go and find it across five systems?
Two, the family structure. Is the data organised at the level of the individual client, or at the level of the family? Because wealth passes down through families, not through isolated bank accounts.
Three, what's off the balance sheet. Does the institution only see what it manages, or does it also see the rest: the real estate, the life insurance held elsewhere, the family business, the accounts abroad? That is exactly the challenge behind a data model built at European scale.
Four, freshness. Is the data updated regularly by the client and the advisor, or is it stuck in a PDF report from 2021?
If the answer to two of these questions is "no," AI won't produce useful results. It will produce statistically clean hallucinations.
What it costs to skip this step
Teams that rush into AI without fixing the data all end up in the same place.
Pilots that give convincing results on two clean files. Then a rollout that collapses on the third, because the other clients' data doesn't follow the same format. Advisors who lose trust. A budget spent with no visible ROI. And an IT team that spends six months cleaning up what should have been clean from the start.
It's a scenario we see everywhere. Not because the teams are bad. Because the order of priorities was reversed.
The right order to do it in
The sequence that works comes down to three steps.
First, reconcile. Everything that belongs to a client and their family should come together in one place, with a consistent structure. Bank account, real estate, a stake in a company, life insurance, a past gift. All of it.
Then, add context. A single line of wealth says nothing on its own. "An apartment in Brussels" is worth little. "An apartment in Brussels held in joint ownership by the husband and the children from a first marriage" is worth a lot. Context is what makes advice possible.
Finally, make it readable. For the client and for the advisor. Because data that stays in the belly of the system produces no value.
Only at that point does it become worth putting AI on top. To extract documents, summarise, suggest the points to raise. Never before.
What it changes for advisors
When the data is in order, the advisor stops losing time on tasks that aren't theirs to do.
They stop re-keying figures. They stop hunting for where the notarial deed is. They stop recalculating net worth by hand. That time, which doesn't exist in a normal day, comes back. It goes back into the client conversation. Into the questions that really matter: protecting the spouse, passing on wealth, the family project.
AI doesn't create that time. It reveals it, once the data has freed it up. That's what AI changes for advisors when the foundations are sound.
The question to ask before signing an AI project
If you lead the technology strategy of a private bank or an advisor network, there's one question that frames things very fast.
"If I take the AI out of the project, does the rest still have value?"
If yes, you're on a serious project, one that deals with the data, the process, the relationship. AI will accelerate it, not save it. It's also the moment to rethink the wealth planning offering in private banking.
If no, you're on a communications project. And it will probably cost more than planned.
