Bank Fee Analysis: How AI Agents Catch Overcharges

Bank fees are one of the last places in finance where money leaks quietly and almost no one looks. Every month, each bank sends an account analysis statement, a dense breakdown of per-service charges priced against AFP service codes, and for most companies it goes straight into a folder. Rates drift from what was negotiated, services get added, and volumes change, and unless someone reconciles the statement line by line against the agreement, overcharges just clear.

Bank fee analysis is the discipline of catching that, and it is tedious enough that it rarely happens consistently. It is also a near-perfect job for an AI agent, because it is pure comparison at scale: parse the statement, match each charge to the negotiated rate, and flag what does not agree.

What bank fee analysis is

Bank fee analysis is the process of checking the fees a bank charges, itemized in the monthly account analysis statement, against negotiated pricing and the services actually used, to catch overcharges, billing errors, and drifting rates. It relies on standardized AFP service codes, and done manually across many banks and accounts, it is slow enough that most teams skip it.

Why bank fees go unchecked

The scale of the problem is well documented. EY's corporate research found that more than 80% of companies do not periodically reconcile their bank charges, more than 60% are dissatisfied with the clarity of those charges, and companies that do analyze their fees typically achieve 20 to 30% savings. The reason it goes unchecked is the statement itself: the account analysis statement is built for banks, not for easy review, with hundreds of line items across service codes, per account, per bank, every month. Reconciling it means mapping each charge to the negotiated rate card, checking volumes, and spotting the service that was quietly added or the rate that crept up. Across a dozen banking relationships that is hours of matching no one has, so it slips, and the leakage compounds because an unchallenged overcharge simply recurs.

What AI agents do with bank fees

• Parse the statements. Read the monthly account analysis statements across every bank and account into a normalized view by service code.
• Compare to negotiated rates. Match each charge against the agreed rate card and the services you actually use.
• Flag discrepancies. Surface overcharges, rate drift, duplicate or unexpected charges, and services billed but not used.
• Track across banks and time. Roll fees up across relationships and periods so trends and new charges are obvious.
• Prepare the case. Assemble the detail to take back to the bank for a credit, with everything traceable to the statement.

How Concourse runs bank fee analysis

Bank fee and account analysis is one of the treasury workflows Concourse agents run. They connect to your banking data, parse the account analysis statements, compare charges to your negotiated rates, and flag discrepancies for review, with every flag traceable back to the source statement. It is SOC 2 Type II certified and keeps a person approving before anything goes back to the bank. This is part of the cash-desk work in the AI treasury analyst.

Frequently asked questions

What is bank fee analysis?

It is the review of a bank's monthly account analysis statement against negotiated pricing to catch overcharges, billing errors, and rate drift. It uses standardized AFP service codes to compare charges to the agreed rate card across accounts and banks.

How do AI agents help with bank fees?

Agents parse the account analysis statements, compare each charge to your negotiated rates and actual usage, and flag overcharges and discrepancies across every bank and period, turning a slow manual reconciliation into a continuous check, with a person approving any claim back to the bank.

Is bank fee analysis worth doing?

For companies with meaningful transaction volume across several banks, yes. Billing errors and rate drift are common, and because an unchallenged charge recurs every month, catching it once pays off repeatedly. The barrier has always been the manual effort, which is what automation removes.

The bottom line

Bank fees leak because the statements are tedious and the review is optional, so it slips. An AI agent removes the effort: it reads every account analysis statement, checks each charge against what you negotiated, and flags what does not match, turning an annual scramble, if it happens at all, into a standing control.

If you want agents watching your bank fees against your negotiated rates, talk to our team.