The Fraud Control That Was Never on Anyone's Budget

Roughly fifteen percent of American bank branches disappeared over the decade through 2024, and the pace has not meaningfully let up, with closures running in the low hundreds in the first quarter of this year alone. The explanation offered is almost always the same and is almost always accurate: customers moved to mobile banking, foot traffic fell, and overlapping locations from a merger cannot all be justified. Every part of that reasoning holds up. What it leaves out is that the branch was doing a second job nobody ever put on the ledger, and that job is now being eliminated by accident.

Bank branch employees have become one of the more effective interception points for elder financial fraud in the country. An older customer arriving to wire a large sum, empty a certificate of deposit early, or withdraw an unusual amount of cash presents a pattern that trained front-line staff are taught to recognize, and the intervention happens face to face, in the narrow window before the money is gone. More than a thousand financial institutions now run their staff through fraud recognition training built for exactly this situation, and at least one large bank has built dedicated scam interruption teams with input from behavioral scientists. That infrastructure exists because it works. It also depends entirely on the customer physically being in a room with someone.

Sponsored by TradeSmith

This is Where The Next AI Millions Could Be Made

Dear Reader,

Billionaires are fleeing the country...

Building million-dollar bunkers...

And hoarding guns, gold and gas masks.

What in the world do they know that you don't?

An even better question:

What are you going to do?

I know why they're running scared, and I'll tell you everything - including what stocks you need to buy by August 31st to protect yourself.

The stocks aren't Nvidia, or SpaceX, or any of the Mag 7.

As a matter of fact, you don't even have to invest in any AI stocks.

I've spent $17 million on AI tools and research, work with a dedicated team of data scientists and even hired a former NSA codebreaker and Pentagon insider to help develop these tools.

And I can tell you - There's a better way to profit from AI, and it's not at all what you think.

That's why today I'm going to show you where I believe the next AI millions are most likely to be made.

But you have to move fast...

August 31st is right around the corner, and you need to be ready to make this move to avoid the consequences that are coming.

Click here to get all the details.

Best,

Keith Kaplan,
CEO, TradeSmith

P.S. Billionaires are hoarding supplies of guns, gold and gas masks. They already know what's coming and they're getting ready. You need to prepare too - Here's everything you need to know.

This ad is sent on behalf of TradeSmith at 1125 N. Charles Street, Baltimore, Maryland 21201. If you’re not interested in this opportunity, please click here.

What Makes the Counter Different From the App

The reason a screen cannot substitute here has to do with the nature of the fraud rather than the technology. In the scams that drain retirement accounts, the victim is not tricked into clicking something. They are persuaded, over days or weeks, by someone who has built a relationship and coached them on exactly what to say if anyone at the bank asks questions. By the time the transfer happens the victim believes the transaction is legitimate and often believes that anyone trying to stop it is the actual threat. An automated fraud flag cannot work through that. A conversation sometimes can, because the person across the counter can notice hesitation, an inconsistent story, a phone call the customer keeps stepping away to take, and can slow the transaction down long enough for the spell to break.

The Mechanics of a Branch Decision That Never Prices the Teller

Here is how a closure actually gets decided, and why this externality is invisible in it. A branch is evaluated on deposits held, transaction volume, loan originations, and the cost of the lease and staff. In a merger, two banks with overlapping footprints run the same analysis across the combined map and close whichever location is redundant. Every input to that calculation is a revenue or cost line that belongs to the bank. Fraud prevented at the counter belongs on a different ledger entirely: it is a loss avoided by the customer, in a transaction that, having been stopped, generates no record of ever having been at risk. The bank does not book it, the regulator does not require it to be counted, and the branch analysis has no field for it. So the closure decision is made correctly on every metric that exists, and the thing that has no metric is what gets removed. This is a common shape in consolidation generally, where the functions that vanish first are the ones no line item was ever assigned to.

Who Loses the Room

The distribution matters as much as the total. Branch closures concentrate in rural counties and in lower-income urban neighborhoods, precisely where the remaining customers are older, more likely to bank in person by preference, and least likely to have an adult child nearby who reviews their accounts. A customer who now drives forty minutes to a branch does not make that trip for a routine question, which removes the incidental contact where a teller might notice that something has changed. Increasingly the alternative is a call center, where the same training can be applied but the cues are thinner, or an app, where there is no interception at all and the transfer completes in seconds.

What Replaces It, and What Does Not

Banks are not indifferent to this, and the investment in call-center scam interruption is real. But a fraud control that was never counted as a service cannot be replaced deliberately, because no one measured what the branch was providing until it was already gone. For a household thinking about this practically, the useful arrangements are the ones that reconstruct a second pair of eyes: naming a trusted contact on the account, which most banks now offer and which lets them call someone if a transaction looks wrong, setting up alerts on large withdrawals, and knowing before it matters whether the remaining local branch still has staff who know the customer by name.

— John Stone