SMA Crown Confidential Notebook – Note: HSBC's Trust Threshold

Occasion: HSBC’s “The Trust Threshold” survey (~10,000 affluent/HNW investors, 10 markets, Ipsos-fielded) — 73% use AI for finance, but the most valued human functions are applying judgment and validation (32%), catching mistakes in AI-generated data (29%), and personalised interpretation (28%); investors delegate research but want a person to hold the decision.

The position occasioned: the question above the question — foregrounded as the lead for the first time, rotating off the side-of-table and person-not-assets angles that led the last several weeks. The investors have drawn a boundary at the decision layer; the position notes the quieter layer beneath it that no survey asks about.

A survey published this week put the question of AI and trust to some ten thousand affluent investors across ten markets, and found them drawing a careful line. Most now use AI to gather and inform. But the functions they most want kept in human hands are judgment and validation, catching the mistakes in what the machine produces, and interpretation made personal to them. They will delegate the research. They want a person to hold the decision.

The finding comes from a wealth manager, and a wealth manager’s survey concluding that clients want wealth managers is worth reading with that in mind. But the instinct the investors report is sound on its own terms, and it is more interesting than the use the finding will be put to. It is worth following one layer further than the survey does.

The judgment these investors want kept human is judgment about the decision — what to do with the holdings, which move to make, whether the machine’s output can be trusted. That is a real boundary, and a wise one. But it is not the deepest boundary. Beneath every decision about what to do with wealth sits a prior question the survey does not ask, because the sector is not built to work there: not what to do with what you hold, but who to be with it. What the wealth is for. What the person holding it intends to become through the holding.

That is the layer where being understood personally actually lives. Validating a data set, correcting an AI’s error, interpreting a chart in light of a portfolio — these are valuable, and they are all performed on the assets. None of them reaches the person. To be understood at the level beneath the decision, an intelligence has to be formed from that person to begin with — from their history with what they hold and the questions they keep returning to — rather than fitted to them, however skilfully, from a model built out of many.

The investors have located the right instinct. Keep the judgment human. We would only follow it one step down: the judgment worth protecting most is not the judgment about the decision, but the judgment about who you are underneath it — and that is a question no survey, and no intelligence formed from the market rather than from you, is positioned to hold.

Seraphina Michelina Evelyna Bogomolova-Huotelin

Founder & CEO of SMA Crown Confidential

Digital Confidantes: Bespoke AI Intelligence for Private Decision-Makers