The Muse Selloff Prices the Part of Advice an App Can Do

Charles Schwab, Ameriprise Financial and Raymond James fell together in Tuesday's session, and the catalyst was not an earnings miss or a rate print. It was an app. Meta's Muse agent has climbed to the top of Apple's free-app rankings, and according to AdvisorHub's morning brief, the declines came as investors weighed how AI agents could change the economics of financial advice and distribution. The flinch lands on a market already primed to move: long-dated G7 government bonds have lost roughly half their value since 2020, and the iShares U.S. Treasury Bond ETF is down 3.6% this year, per the same wrap.

Read the capabilities behind the ranking twice. Muse already handles tasks that run from travel bookings to transactions, and the market's logic continues from there: if a consumer agent will compare products, execute transactions and absorb the routine chores, then the front door of the advice relationship stops belonging to the firm.

Note which names took the hit. A custodian and two distributors sold off in the same session, all of them sitting between the client and the product shelf, and that is exactly the space a consumer agent is built to occupy. Schwab's referral engine and product menu, Raymond James' recruiting and distribution machine, Ameriprise's brand-led funnel — each sits in the path Muse is angling to walk. That reading of the tape is an inference, but it is the one that fits the price action.

Strip the fear out and a concrete problem is left on the desk. The transactional layer Muse is being priced to eat — comparison, execution, small-task servicing — is the layer many practices still bill for, because it generates activity and, historically, the first phone call. A practice whose growth model is answering whatever the client asks is now competing with a free app that answers faster and never loses the file. The defense is not a speech about the value of human advice. It is a narrower service catalog, and a harder conversation about which everyday questions now have a better home than the one they used to have here.

This is the split we have been tracking, and it moved again this week. The AI adoption gap has become a supervision gap: the advisor who cannot document what the loop actually does is already behind, and, as this publication argued when Anthropic's advisor plug-in arrived, a connector is worth signing only for the firm that already knows what the work cost before it showed up. Signing the tool is the cheap part. Knowing what it does without you, and being able to show it, is the part that survives scrutiny.

Client acquisition bites first, and the precedent is set. Schwab ends sub-$5 million referrals to firms in its Advisor Network starting in 2027, a floor that pushed network RIAs to build their own pipelines — a change this publication reported when the cutoff was announced. If a consumer agent becomes the default first stop for comparing products, the referral math thins again, and the practice that never owned its own lead generation loses the only engine it had.

Tuesday's tape is one session, and no practice should underwrite a decade off a chart. But the brief's own framing is the tell: AI disruption is being reclassified from a productivity story to a competitive threat. That reframing does more damage to a growth plan than any down day, because it changes what a client believes they are paying for. The advisor who can name the human deliverable in the next meeting is ahead of the one still selling access and activity.

If the transactional layer is software, growth has to come from the layer above it, and that shelf is filling. Blackstone, KKR, EQT and other managers have attracted roughly $100 billion into perpetual private-equity funds, and U.S. perpetual equity assets have roughly doubled since early 2025, according to the same brief. Performance across those vehicles varies widely, and some lean hard on soaring private-technology valuations — which is precisely the due-diligence problem a practice inherits when it puts them in front of a wealthy client. A perpetual structure is built for small tickets; deciding what belongs in it, and being answerable for that call, is work an agent cannot do for the client.

Meanwhile the recruiting market keeps its own clock. Raymond James, whose stock fell Tuesday, has been competing on client experience rather than payout — the firm recently brought in a Schwab custody veteran for exactly that, a hire this publication read as a test of whether relationship quality decides the next platform contest. Per WAD's records, Raymond James also booked a $425 million team liftout and a three-advisor team in the same stretch. The next move will go to the firm that can show the client experience the platform delivers, because that is the part of the practice the app cannot take with it.

The selloff's verdict — that an app can handle the comparison and the transaction — is probably right. What it does not answer is who owns the part that comes after. The first compliance exam that asks a firm to map its own AI loop will separate the practices that adopted the tool from the ones that understood it, and it will do that with more finality than any download chart.