Hacked email or a fake invoice: what to do straight away, in the first hour
A supplier announces a new IBAN, the director urgently asks for a payment, and colleagues receive odd messages from your address. What to do straight away, whom to report to, and how to set rules so that a payment to someone else's account does not happen.
The most expensive fraud in small and medium-sized companies usually does not look like an attack. It looks like an ordinary email. A supplier says they have changed banks, and the attachment is an invoice with a new IBAN. The director, travelling, writes that an advance must be paid urgently and asks not to be called because they are in a meeting. The message comes from the real address, or from one that differs by a single letter.
The money goes, and the fraud comes to light a week later, when the real supplier sends a reminder.
This article has two parts: what to do straight away when it happens, and how to prevent it.
The first hour: if the money has already been paid
Speed matters most. A payment can sometimes be stopped or recovered only in the first few hours.
- Call your bank straight away. Do not email: call the number you already have, not one from the suspicious message. Say it is fraud and ask for the order to be recalled or stopped. Write down the person’s name, the time of the call and the reference number.
- Report it to the police. A criminal complaint for computer fraud is often a condition for the recipient’s bank to hold the funds. Bring printouts of the messages, the invoice with the fake IBAN and the payment confirmation.
- Tell the real supplier, on the phone number you had before, not using details from the fake message. Their email may have been hacked rather than yours, which puts their other customers at risk too.
- Keep the evidence. Do not delete messages. Forward them as attachments, with full headers, to whoever will investigate.
The first hour: if the email account has been hacked
The signs are messages you did not send, replies to emails you never saw, colleagues asking “did you send this?” and sign-ins from unfamiliar countries.
- Change the password for that account, from a computer you know is clean. The new password must be long and new, not the old one with a changed digit.
- Sign out all sessions in the account settings. Otherwise an attacker who is already signed in stays signed in after the password change.
- Turn on two-step sign-in, if it is not already on.
- Check the mailbox rules. Attackers often set up a rule that forwards messages to an outside address or deletes them automatically, so the owner does not see them. Check forwarding and automatic replies too.
- Warn customers and suppliers not to trust messages from that address about changed payment details during this period.
- Change passwords elsewhere wherever the same password was used.
Whom to report to
| To whom | When |
|---|---|
| The bank | straight away, if money was paid |
| The police | for any fraud involving money, and for attempts |
| The National CERT ([email protected]) | a computer security incident, including phishing and a hacked mailbox |
| The Personal Data Protection Agency (AZOP) | if personal data were accessible to the attacker, within 72 hours of becoming aware, unless the breach is unlikely to pose a risk to people |
| The competent CSIRT | if the company is subject to the Cybersecurity Act, according to that act’s deadlines |
For companies subject to the Cybersecurity Act, deadlines and the reporting procedure are described in the article NIS2 and the Croatian Cybersecurity Act.
A hacked mailbox almost always contains personal data: names, addresses, contracts, sometimes payroll data. That is why the question of reporting to AZOP should be asked straight away, not once things have calmed down.
How to recognise a fake message
- A change to payment details. Any message that changes the IBAN, the bank or the payee is suspicious until confirmed by phone.
- Urgency and secrecy. “By 2 pm today”, “don’t tell anyone”, “I’m in a meeting, don’t call”.
- An address that is almost the same. finexa-hr.com instead of finexa.hr, rn instead of m, a zero instead of the letter o. On a phone, the address is often not even visible.
- A reply in an old thread. An attacker who has been reading the mailbox writes into an existing conversation, with real names and quote numbers. A genuine thread is no proof that the message is genuine.
- An attachment that asks you to sign in. “The invoice is in a shared document, please sign in” is the most common way passwords are stolen.
Rules that prevent the damage
Technology helps, but what protects the money is an agreement within the company.
- An IBAN change is confirmed by phone, on the number from the contract or earlier correspondence, never on a number from the message asking for the change. No exceptions, even when the message comes from the director.
- Two people for larger payments. One prepares the order, the other approves it. Most online banking supports this.
- A new payee is checked before the first payment. The account holder’s name must match the supplier.
- Two-step sign-in on all email accounts, especially those of management and accounting.
- Domain protection. SPF, DKIM and DMARC records for your domain make it harder to send messages impersonating your address. It is a one-off setup.
- Short training, twice a year. Half an hour with examples of real messages is worth more than a long manual.
- Backups an attacker cannot delete. When ransomware arrives along with the email, a copy kept separate from the network will save you. More in the article on the 3-2-1 rule.
E-invoices reduce one risk, but not all of them
Since 2026, VAT-registered companies in Croatia exchange e-invoices through intermediaries, and from 2027 traders and companies outside the VAT system do the same. A fake PDF invoice sent by email is therefore easier to spot, because the real invoice arrives by a different route. A “we have changed banks” message can still arrive by email. The rule about confirming by phone is still needed. More in the article on Fiscalisation 2.0 and e-invoicing from 2027.
Frequently asked questions
Can the money be recovered? Sometimes, if the bank is told straight away and the money has not yet been withdrawn from the recipient’s account. The more time passes, the lower the chances.
Who bears the loss if we paid to a fake IBAN? As a rule, the company that paid, unless shown otherwise. The debt to the real supplier usually remains. That is why the phone confirmation rule is cheaper than any insurance.
Is changing the password enough? No. You need to sign out all sessions, check forwarding rules and turn on two-step sign-in. Otherwise the attacker stays in the mailbox.
Do we have to report to AZOP? If personal data were accessible to the attacker, probably yes, within 72 hours. Write down the decision and the reasons in any case.
If something has just happened, or you want to check how exposed your company is, we do that as part of cybersecurity and the Cyber-Fin Security review, which looks at both the technology and the flow of money. Get in touch.