Edelman skipped the Claude launch table, and that is the tell

Anthropic's launch of Claude for Financial Advisors was, by InvestmentNews's account, the big news at this year's Future Proof festival, and the most revealing thing about it is a firm that was not on the partner list. Charles Schwab, Vanguard and Dynasty Financial were among the wealth management players Anthropic announced alongside the product; Edelman Financial Engines, with $326.3 billion in regulatory assets and 1.5 million accounts, was not among them, and its chief executive, Ralph Haberli, told InvestmentNews the firm has been working closely with Anthropic for about the last six months anyway.

A partner list is a channel announcement; a direct relationship is an integration project, and the two are priced very differently for the firm that signs — when Schwab took its Claude seat, this publication's read was that the seat buys queue position rather than a better model, and nothing from Huntington Beach disturbs it. What changes a planner's week happens below the logo — the recording of the client meeting, the preparation that precedes it, the workflow that follows — and it gets assembled one integration at a time.

About 5,700 attendees traveled to Huntington Beach, Calif., for the festival, and AI was the topic that followed them into every room. Haberli called it 'a revolution in many respects,' pointing to 'a ton of excitement' and 'a ton of new capabilities,' while conceding that the same technology can be confusing and disconcerting for the advisors expected to use it. The coverage's own framing is of an industry shifting gears in its use of AI — from curiosity about the technology to deployment inside the work advisors already do — and the partner roster was the clearest evidence of that turn. Both halves of Haberli's description are worth taking at face value, and the second half is the more interesting one for anyone selling into advisory firms.

His sharpest observation was about clients, who increasingly arrive having done their own research — 'they've gotten the view on asset allocation, tax planning from Claude or ChatGPT or others' — and he compared that to the WebMD search everyone used to run before a doctor's appointment, the patient walking in with a diagnosis the physician has to examine rather than accept. Read the comparison as an argument about where advice fees survive. A client who arrives with an allocation and a tax view has already absorbed the search cost that advice used to price, and what remains to charge for is the examination: the part of the job that requires looking at the household in front of you.

Edelman's own deployment is deliberately unglamorous: Haberli described AI that records client meetings, helps prepare for them and shortens the workflow around them, framing the firm's test as tangible impact rather than technology adopted for its own sake. Multiple of Edelman's best planners, he said, report that how they work is 'radically different and radically better' than it was six months ago. 'We're very much excited about the impacts of it,' he said, 'but we're trying to keep it tangible and real, and we can point to the before and after.' That before-and-after is the only AI metric a buyer can act on, and it does not fit on a launch slide.

The announced partners described a different bargain: Eve Cout, head of advisor solutions in Vanguard's financial advisor services unit, said the point of working with Anthropic is to put Vanguard's asset allocation and portfolio construction in front of advisors where they already are — the firm's models, she said, are used by advisors as a starting point as they build portfolios, and 'if they're going to do it in Anthropic, in Claude, we want to make sure that information's there and they can just easily access it.' That is a distribution statement, and a rational one. Where models are hard to tell apart, placement is the lever that remains, and the interface where an advisor starts the morning is the newest shelf in the store.

It also lands on the constraint at the center of this month's coverage: the client record is the expensive part of AI. A general-purpose model is available to anyone who signs a partnership; a governed, permissioned copy of what the client said, held and was told to do is the piece a model vendor cannot hand over in an announcement. Edelman's arrangement reads as a bet that the integration is the asset and the announcement is advertising — a bet a firm of its size can afford.

The arithmetic supports the reading: PWD's records put Edelman Financial Engines at 1,679 employees against 1.5 million accounts and $326.3 billion in regulatory assets, which works out to roughly 900 accounts and $194 million of regulatory assets per employee. At that density, meeting capture and meeting preparation stop being novelties and become the only way to hold a service model steady while the book grows. A firm that can build that plumbing keeps the difference; a firm that cannot will rent it, and rents get repriced. The mid-market RIA taking a vendor's template this fall is making a sensible call on cost and a real one on differentiation, since whatever the template does well it will do the same way for the firm down the street.

Haberli's planners can already point at a before-and-after, and the 5,700 attendees who heard the launch announcement can judge for themselves how many partner firms will be able to say the same a year from now.