Supplier bank account change fraud: how not to lose 200,000 on a single email

· 7 min read · SilverData team

A supplier you've worked with for five years sends an invoice like every month. Same email address, same signature, same tone. Only the bank account number in the message is different — "please pay to the new account, the previous one is being closed". The transfer goes out on Friday afternoon. On Monday the real supplier calls asking when you're going to pay. That's when you realise the money went to someone else entirely.

This isn't a hypothetical scenario. It's the most expensive attack on SMEs today — business email compromise — and unlike phishing with a crooked bank logo, nothing here looks suspicious. In one of our projects, a 30-person construction subcontractor from Silesia lost 187,000 PLN on a single transfer like this. They recovered 62,000 PLN after months of correspondence with banks and police. The rest was gone.

This post isn't about "training people to be vigilant". Vigilance doesn't scale. This post is about how to build a simple procedure that stops even a tired Friday-afternoon bookkeeper from sending money to the wrong account.

Anatomy of the attack

The modern version of this scam usually goes like this. The attacker breaks into your supplier's mailbox — usually via a stolen password or phishing, no MFA. For a few weeks they only read. They learn how the supplier writes, how they sign emails, when invoices go out, who on your side approves them. They already know invoices arrive on the 25th to Anna in accounting, and Anna doesn't call back — she just books.

When the moment arrives, the attacker sends a real invoice from the real mailbox, with one change: the bank account number. Sometimes swapped in the email body, sometimes in the PDF. Sometimes they send a polite "heads-up" a week earlier — "we're notifying you of a change of bank account" — on company letterhead, with a scan of the owner's signature. The invoice with the new account that follows looks entirely natural.

There's no suspicious domain. No weird attachment. SPF, DKIM, DMARC — all pass, because the email really does leave the supplier's server. Anti-phishing filters won't react. That's why the attack works so well: technically everything is fine, only reality has been swapped underneath.

Why SME processes fail here

At a large corporation, changing a supplier's payment details goes through procurement, verification in a vendor register, and manager approval. At a 20–40 person SME, that chain is usually missing. The invoice reaches the bookkeeper by email, the bookkeeper books it, the owner signs the transfer in online banking — often without looking at the account number, because "it's from Nowak, we've been paying him for five years".

Three assumptions that fall apart in practice:

You're not missing a "detection tool". You're missing one step in the process: independently verifying the bank account change through a channel different from the email that requested it.

The "callback" procedure: one rule you write into your policy

The only rule that works, and that the attacker can't route around, sounds like this: every change of a supplier's bank account must be confirmed by phone, on a number from the vendor record in your ERP — not from the signature in the email.

For a 20–50 person company, in practice it looks like this:

Two supporting mechanics:

That's it. You don't need a SOC, an EDR or an expensive "anti-BEC" tool. You need one sentence in your policy and a "bank account confirmed by" field in the supplier record.

The first 30 minutes after detection

Suppose the transfer went out anyway. Every minute matters — literally. The money usually lands in a "mule" account and is withdrawn or forwarded within hours. Order of operations:

  1. Call your own bank's business hotline and request a transfer recall. Get the case number. For SEPA and domestic transfers the chance drops with every hour — after 24 hours it's a few percent.
  2. Contact the recipient's bank (you'll see it on the transfer confirmation) — directly, asking them to freeze the funds on the account.
  3. File a police report — the duty prosecutor can issue an order to secure funds faster than a routine case would. Get the case number.
  4. Report to CSIRT NASK (incydent.cert.pl) — even if it's "just" money, the incident gets correlated with other reports and can help trace the whole campaign.
  5. Notify the real supplier — their mailbox is compromised, every other customer of theirs is at risk. It's not a fun call, but a necessary one.
  6. Rotate passwords on your own accounting and management mailboxes and switch on MFA anywhere it isn't on yet. If the attacker reached your supplier, the same route could reach you.

At that Silesian company, a fast recall through the bank and a police report within 3 hours froze part of the money in the mule account before it was withdrawn. The rest was already abroad and gone. Without that first hour, the recovered 62,000 PLN wouldn't have come back either.

A short checklist for this week

You don't need a quarter-long project to close this risk. In a week, the minimum you can do is:

That's a few hours of work. One transfer not sent to a fraudster pays that cost back many times over in a year.

What next

Payment fraud against suppliers isn't a problem that's about to disappear — quite the opposite, in the last two years these attacks have clearly accelerated. Attackers noticed that SMEs move enough cash to lose hundreds of thousands on a single transfer, and have processes simple enough to let it through.

If you want to check what your payment procedure actually looks like and where the holes are — instead of waiting for a fake "Nowak" to reach out — get a second pair of eyes on it. The first call is free and takes half an hour.

Vigilance doesn't scale. What scales is one field on the supplier record: "bank account confirmed by phone by".

← Back to blog