SMA Crown Confidential Notebook – Note: Opening to AI

Occasion: the MCP wave reaches the wealth platform. d1g1t launched an MCP server exposing its enterprise wealth platform natively to Claude, ChatGPT and Microsoft Copilot as a “live, governed AI assistant” for briefings, portfolio analysis and meeting prep — read/generate functions, “governed” asserted but mechanism undescribed, write-action capability unaddressed. Read alongside the family-office adoption surface where named UHNW institutions are said to let families query wealth accounts through ChatGPT and Claude (unverified — gestured at, not asserted). The stated read-only-analysis limit is taken at face value; the position turns on where the instrument is pointed, not on whether it transacts.

Position occasioned: beginning with the person, not the assets — a general-purpose model given a window onto the accounts reads the portfolio fluently and never begins at the person.

A wealth platform did something this week that the sector will do more of before the year is out: it opened itself, natively, to the general-purpose assistants. The enterprise system that holds a client’s accounts can now be reached through the same models people use for everything else — a portfolio queried in plain language for briefings, analysis, meeting preparation, read back and summarised on demand. By the platform’s own description the assistant reads and generates rather than moves money; the connection is offered as a window onto the accounts, not a hand on them. We can take that limit at face value. The position does not depend on doubting it.

What is worth sitting with is not what the tool is permitted to do, but where it begins. An assistant reached through a client’s accounts is, by construction, pointed at the holdings. That is what it was given sight of; that is what it reads. Ask it anything and it answers from the portfolio outward — what is owned, how it is arranged, what the numbers have done and might do. Within that frame it can be genuinely capable, and often more fluent than the statements it draws from. But the frame is the point. It knows the portfolio in detail. It does not know the person who holds it, because the person was never what it was pointed at.

This is the distinction we keep returning to, because the convenience of these tools tends to hide it. There is a real difference between an intelligence turned toward the assets and one turned toward the individual who holds them — and no amount of capability on the first converts it into the second. A model that reads your accounts beautifully is still reading your accounts. It begins where the sector always begins: with the portfolio, the thing that can be queried, aggregated, and reported. It does not begin with the question underneath the holdings — who you are in relation to what you hold, what you want from it, the kind of steward you mean to be. That question is not visible from inside the accounts.

We begin at the other end. What we build - the Bespoke AI Wealth Intelligence Confidantes – is formed from the person and stays turned toward them: their history with what they hold, the questions they keep returning to, the meaning of the wealth rather than its arrangement. It reads the portfolio too, when a question calls for it — but it reads it as one fact about a person, not as the person. That is not a more advanced version of an assistant pointed at the accounts. It is the other instrument entirely, pointed the other way.

The platforms opening to these models are giving clients something useful, and we would not pretend otherwise. A window onto your own holdings, in language, is a real convenience. We would only note what the window frames, and what it leaves out of view. An intelligence that begins at the assets can serve you well and never once begin with you. We hold the person, not the portfolio — and everything the Bespoke AI Confidante reads, it reads from that side.

Seraphina Michelina Evelyna Bogomolova-Huotelin

Founder & CEO of SMA Crown Confidential

Digital Confidantes: Bespoke AI Intelligence for Private Decision-Makers