Protect//guide // cybersecurity
Invoice fraud with a changed IBAN: how money leaves without anyone noticing
The attacker gets into a mailbox, waits for a real invoice and changes only the bank account. How it works, what to check in your company and what to do in the first hours.
Updated: 15.11.2026 · ~6 min read
In its annual report for 2025, Romania’s National Cyber Security Directorate (DNSC) describes an industrial company with 375 employees whose attackers took over two email addresses: the general one and the purchasing department’s. They set up mail forwarding and altered documents inside ongoing email conversations. The result was unauthorised bank transfers totalling 20,148 euros, made by the company itself, into the attackers’ accounts. The case is in the DNSC report (page 13, in Romanian) and was covered by StartupCafe in August 2026.
The report does not go into each transfer, but the mechanism is the usual one in this kind of fraud. The attacker takes a real document from a conversation with a known partner and changes one detail, usually the IBAN. There is no link to click and nothing in the message looks suspicious.
How the attacker gets into the mailbox
Most often, with the password. A fake page that mimics the Microsoft 365 or Google sign-in screen, a password reused on another site that was breached, or an infostealer that copies the passwords saved in the browser. In the case above, DNSC found many such programs on the company’s systems. The report also notes that account compromise incidents in 2025 were more than four times higher than in 2024, and lists password reuse and missing or misconfigured multi-factor authentication among the causes.
What happens once they are in
Nothing is sent right away. The attacker reads the mail for days or weeks and learns who pays whom, how much and on what terms. Then they create a rule that forwards messages containing “invoice”, “payment” or “IBAN” to an outside address, or moves them to a folder nobody opens, such as Archive or Deleted Items. That way they see the replies before you do.
When a large payment is due, they drop the real invoice into the conversation with one line changed, or a message “from accounts” saying the company has moved to another bank. Sometimes they write from the compromised address, sometimes from a domain that differs from the supplier’s by a single letter.
How fast the money goes
At the Orange Business conference on 21 September 2026, the owner of a Romanian transport company said attackers took more than 100,000 euros in about 30 minutes, through 11 transfers, while the company’s phones were flooded with calls so they could not communicate. The Mobilissimo article does not say how the attackers got in, so this was not necessarily invoice fraud. It does show how little time is left once the money starts moving.
What to check in your company
- MFA on every email account, including shared addresses (office@, accounts@) and the administrator account. One account without MFA is enough;
- the rules in each mailbox: in Outlook, under Settings, Mail, Rules. A rule nobody created that sends messages outside or deletes them is a clear sign of compromise;
- automatic forwarding to external addresses. In Microsoft 365 the administrator can block it for the whole company and allow it only where it is really needed;
- the sign-in log: logins from countries where you have nobody, or at hours when nobody works;
- who can change supplier bank details in the accounting software or ERP, and whether the change leaves a trace.
The rule for any new IBAN
A change to a supplier’s bank account is confirmed by phone, on a number you already have in the contract or the company address book, never the one in the email announcing the change. It takes two minutes. The rule applies to the supplier you have worked with for ten years, to the small invoice and to the message marked “urgent”. Urgency is the attacker’s main tool.
Also useful: payments above a set threshold approved by two people, a list of confirmed IBANs for your main suppliers and, following the transport company’s example, a phone used only for the banking app.
What to do in the first hours
- First, call the bank and ask for the transfer to be stopped or recalled. Every hour counts, because attackers move the money on quickly.
- Change the password of the compromised account, sign out all active sessions, delete suspicious rules and turn on MFA if it was missing.
- Warn the partners you have been writing to recently. The attacker may have sent them fake invoices from your address too.
- File a police complaint with the emails, bank statements and the time of each transfer.
- Report the incident to DNSC on 1911, Romania’s national number for cyber incidents. According to the report, the line received more than 4,300 calls in 2025.
Do not wipe the mailbox or reinstall computers before saving the evidence. Sign-in logs, rules and message headers are exactly what the bank and the police will ask for.
MFA, blocking external forwarding and regular reviews of mailbox rules are usually handled as part of a company’s IT management. For a wider look at the risks in your company, the cybersecurity for SMEs page explains what an assessment involves.
FAQ
What companies ask us
Will the bank refund us if we paid into the wrong IBAN?
There is no guarantee. Your company authorised the payment, so the bank is not obliged to refund it. If you call within the first hours, the bank can ask the receiving bank to recall the funds, and whether that works depends on how much of the money is still in that account.
Doesn’t our antivirus protect us against this?
Only partly. Antivirus can stop a program that steals passwords from a computer, but it does nothing against a login with a password that has already been stolen. And the email with the altered invoice contains nothing malicious to detect. Multi-factor authentication (MFA) and a phone check of every new IBAN cover exactly that gap.
We use MFA with SMS codes. Is that enough?
It is much better than nothing. An authenticator app or a hardware key is safer, because some fake login pages also ask for the code and use it immediately. What matters most is that MFA is on for every account, including shared addresses such as office@ or accounts@.
How can I tell whether our company email is already compromised?
The usual signs: forwarding rules nobody created, messages in Sent Items you did not write, replies that never arrive, sign-ins from countries where you have no staff, and partners asking about a new bank account you supposedly announced.
Do you know who has access to your company email?
If you are not sure where MFA, forwarding rules and sign-in logs stand in your Microsoft 365 or Google Workspace, write to us.