The AI wealth fight is now over distribution

CogniCor's announcement on September 13 was a roster: Shannon Eusey, Louis Diamond, Joseph Kuo, Douglas Wilber, Jim Roth, Allison Couch Pratt and Matthew Incitti, all added to its advisory board on the same day, their backgrounds running through advisory, distribution and wealth roles. For a company that sells artificial intelligence into wealth management, the roster is the sale.

The AI contest in this industry has moved off the model: whatever advantage a frontier model gives one advisory firm is available to its competitors within a release or two, and few firms choose a vendor on the strength of what sits underneath the meeting notes. The scarce assets are the introduction, the committee seat, the half hour on the calendar of the person who signs the contract; CogniCor bought seven of them in a day and filed the purchase under governance.

Look at what else the same day traded, and the pattern holds even where money was named: HSBC took two Citi Private Bank leaders twelve days apart—Cayman Wills into the top job at the U.S. private bank and Hannes Hofmann into a new global head of family offices seat—and paid in titles, with no compensation disclosed in the coverage. Citi's answer was to hire Bank of America's Chris Biotti to run its private bank in North America and Bank of America Private Bank advisor Teresa Radzinski, who manages $3 billion.

Only one of those four moves carries a number, and the number belongs to the advisor rather than to the executives. That is no knock on the title hires: a large mandate is how a bank recruits the people who fill it, and creating a family office seat gives HSBC something to offer before the next hire has a name. A seat is a promise about who arrives next, while a $3 billion advisor is a statement about what is already on the books.

Disclosed pay is absent from all four of those moves, and the absence does real work: it leaves the title as the only comparable unit, and a title is the one thing an institution can manufacture without spending cash. A bigger title at a different firm can be worth more or less than the package left behind, and without disclosure nobody outside the two banks can say which. That is convenient for the institutions and useless for anyone trying to rank the day.

The rest of the day's advisor movement ran the same way: a 16-advisor group at WFA Volare left LPL Financial under Matthew Johnson, the largest team move of the day, arriving without an asset figure attached. Eight advisors went from IAM Advisory to Wealthcare Advisory Partners, six moved from UBS to Merrill Lynch, five left Commonwealth Financial Network for &Partners, and Mercer Advisors hired Northern Trust's William Hinson without a book disclosed. The directions vary—bank to RIA, independent to independent, wirehouse to wirehouse—and the disclosure varies more.

A 16-advisor departure is not a small event for any platform, and a move that size would normally be the most quotable item on the tape. The missing asset number is the most instructive detail of the session: the largest headcount move came with the least arithmetic, exactly what you would expect when the selling side has no interest in publishing what it lost and the buying side has every interest in a story about growth.

Sizing the day's moves by disclosed assets would rank them almost at random. Headcount is what arrives on day one; the assets follow through the transition, if they follow at all, and no firm on either side of these moves has a reason to settle the question in public. The $3 billion in the Citi hire stands out for that reason alone—it is the one move of the day where somebody put a figure on the table.

The priced deal flow sat somewhere else entirely: seven closings on the day summed to just under $1.76 billion, and the only wealth transaction among them with a figure on it came to $22 million—a rounding error against the real-asset total and, by the logic of this column, the more consequential number of the two.

That gap is the argument. Capital is still going to buildings and credit, where multiples are legible and the assets sit still. The wealth management transactions that rearrange who owns the client relationship arrive unpriced, because what changes hands is a team's phone list and its standing in a market, and neither fits neatly into a purchase agreement. This publication has made a version of the case about RIA M&A, where Beacon Pointe, SignatureFD, BridgePort and 1834 have all been hiring for the half of the deal that starts after the wire clears—the half where assets get a multiple and the operators who make them worth anything get a title.

The $22 million figure matters less for its size than for its rarity. Wealth management dealmaking rarely produces a headline price to argue about, which leaves the industry's most important transactions—the ones that decide who serves the client—as footnotes under a real estate closing. That is convenient for buyers and maddening for anyone trying to read the tape.

Put CogniCor's seven names back in that frame: advisor-facing software has always been sold a seat at a time, with pricing pinned to users and assets on a platform, which makes a vendor's growth a function of how many firms will take a meeting. CogniCor's roster suggests a different bid—skip the model comparison and buy the people who get invited to the meeting in the first place.

An advisory board is also the cheapest way to buy a reference call: software decisions at advisory firms get made on the recommendation of people the buyers already trust, and a vendor with seven recognizable names on its letterhead skips a step that otherwise costs quarters of selling time. The names do not close anything by themselves; they get the vendor into the room where the decision gets made, which in a category this crowded is most of the work.

If that is the strategy, it is a repricing of the field rather than a product decision: an advisory board is cheap next to a sales force, and it is cheap for a specific reason—it obliges nobody to do anything. No exclusivity, no quota, no disclosed retainer, no guarantee that a single introduction becomes a pilot. That combination is what makes the play copyable within a quarter, as the first-mover advantage lasts until a competitor's head of business development drafts the same list.

Watch for that answer before the spring conference season. One rival roster announcement would settle what this category believes it is selling, and the sharper question is what the next AI wealth vendor gets asked when it raises: model performance, or the names on its board. My read is the second: distribution people are how this category gets paid, and the vendors who work that out first will price their rounds against relationships rather than benchmarks.

There is a version of CogniCor's move that goes wrong, and it deserves saying plainly: seven well-chosen names can be a marketing line item rather than a channel, because advisory boards are better at producing announcements than deployments. The test is narrow and unglamorous: does a client announcement arrive with one of the seven inside it? One advisory firm saying a board member sat in on its decision is the difference between distribution and a press release, and it is the thing to watch over the next two quarters.

If the seven convert into named work, the next AI vendor to raise a round will be asked for a roster instead of a demo, and CogniCor will have bought its distribution at the pre-repricing price.