For decades, banks have counted on one thing: most customers don’t move their money. Checking and savings balances sit at low rates because comparing accounts and moving cash takes effort, and most people never get around to it. That inertia is one of the cheapest funding sources a bank has.
AI agents that manage money on a customer’s behalf could take that effort away. Wall Street is already treating it as a real threat, and bank marketers should too.
What happened this week
Meta released a new AI agent called Muse this week, and it went viral. According to Yahoo Finance, the app can link to a user’s financial accounts, track balances and investments, make recommendations and take actions for the user.
Investors didn’t wait to see how it plays out. The KBW Nasdaq Bank Index dropped about 2.6% on Tuesday, and stocks including JPMorgan Chase and Charles Schwab were caught up in the selloff. Bank stocks have since recovered some of that ground.
Bank of America analyst Ebrahim Poonawala called “deposit sorting,” the frictionless movement of excess cash into higher-yielding places, a real threat to industry margins. He also said the proof will show up in deposit costs. Until those rise faster than rates or competition can explain, he wrote, the disruption thesis “remains conceptual.”
That’s a fair caveat. It’s also not a reason to wait.
Why this hits harder now
The timing is bad for banks that rely on sleepy deposits:
- The rate gap is huge. The national average savings rate is 0.37%, according to FDIC data, while top online savings accounts pay close to 4%. An agent scanning a customer’s accounts will spot that gap instantly.
- Rates are going up, not down. The Fed raised rates on September 16 for the first time since 2023, which adds pressure to pay depositors more.
- Banks need deposits. Loan growth has picked up this year, and the personal saving rate is near a four-year low, so there’s less new cash to go around.
- Customers already switch. In Raisin’s 2026 survey, 65% of Americans said they’ve changed banks at least once. The main thing holding many others back is hassle, and hassle is exactly what an agent removes.
We’ve seen a version of this before. In 2023, as rates climbed, Schwab customers moved billions out of low-rate accounts into money market funds. Schwab had to lean on expensive short-term funding, and its profits fell for the year. That shift was driven by people doing it manually. Software doing it automatically would move faster.
The big banks are already reacting. Citigroup launched a savings rate program this week aimed at pulling in more of its customers’ cash, following similar moves by PNC and Bank of America. JPMorgan’s Jamie Dimon has talked about an upcoming cash product called Smart Cash.
What an agent will “see” when it looks at your bank
Picture a customer who asks an AI assistant to make sure their cash is earning a fair rate. The agent will look at the balance, the rate, what the customer needs for bills, and what’s available elsewhere. It won’t care about your branch, your brand campaign or the free pen at account opening.
If you’re paying 0.05% on a $40,000 savings balance, the agent will recommend moving it. And if the customer has told it to act, it may just do it.
What to do before this shows up in your numbers
Find your exposed balances now. Pull a list of customers with large balances in low-rate checking and savings accounts. That’s the money an agent will go after first, and it’s the money you most need to protect.
Offer the better rate before someone else does. If you have a competitive savings, money market or CD product, bring it to those customers directly. Moving money within your own bank is cheaper for you than losing it.
Make an internal sweep easy. An automatic sweep from checking into a higher-yield account at your own bank answers the agent’s question before it’s asked. The cash stays with you.
Publish your rates clearly. Agents will compare what they can read. If your rates are buried in PDFs or behind a “call us” button, you won’t show up well in the comparison. Keep current rates on a plain, public web page.
Give customers reasons that aren’t rate. Agents optimize for yield, but customers still choose things like a local lender, a banker who answers the phone, or help with a small-business loan. Tie deposit relationships to those services so there’s more at stake than an APY.
Watch outflows closely. Track transfers to outside accounts, especially online banks and brokerages, and flag customers who start moving money in regular chunks. Rising deposit costs are the lagging signal. Outflows come first.
Don’t wait for proof
Poonawala is right that nobody can prove agents are changing behavior yet. But banks that wait for the data will be responding after the money has already moved. The cash your customers haven’t touched in years is the most exposed money on your balance sheet. It makes sense to start protecting it now, while it’s still cheap to do.
