BEC and Fake Invoice Fraud: Technical and Process Controls

Which controls should be set up against fake invoice and executive impersonation fraud? Look-alike domain detection, mailbox rule monitoring and a verification process for payment changes.
Technical summary
Technology
Exchange Online, Defender for Office 365, DMARC, dnstwist, Microsoft Graph PowerShell
Topic
Technical and process controls against business email compromise
Level
Intermediate
Prerequisites
  • Exchange Online PowerShell access
  • Knowledge of finance and purchasing processes
In this article7 sections

No malware in the attachment, no link, no typos. Just a short email that reaches the right person at the right time: “Our bank account has changed; could you make this month’s payment to the new IBAN?” That’s how business email compromise (BEC) works. This guide explains typical BEC scenarios in organizations that run finance and purchasing correspondence by email, the technical controls that catch them, and the payment verification process that no filter can replace.

What sets BEC apart from phishing

Classic phishing aims to get the user to click a link or open an attachment; most defense layers are built around scanning that payload. A BEC message has no payload to scan. The text is short, the language is clean and the request fits the workflow. That’s why URL reputation checks, sandboxes and signature-based filters largely work in vain. A filter reads the text; it can’t read intent.

Feature Phishing BEC
Target Credentials, malware Money transfer, sensitive data
Payload A link or attachment Usually none, text only
Volume Bulk sending A small number of targeted messages
Condition for success A single click A process gap and trust

General phishing defense layers are covered in the reducing phishing risk in corporate email article. Here we focus on the area those layers don’t see.

Typical scenarios: CEO fraud, fake invoices, supplier compromise

Executive impersonation

The attacker writes to the finance team in the name of the general manager or finance director. The display name is right; the address comes from a free mail service or a look-alike domain. The request is both urgent and confidential: “I’m running an acquisition negotiation; you need to make this transfer today without telling anyone.” The information that the executive is traveling or in a meeting has been gathered from the corporate website and social media. The timing isn’t a coincidence.

Fake invoice

The attacker impersonates a real supplier and sends accounting an invoice or an “account change” notice. The invoice template has been copied from the supplier’s website or from previously leaked documents. The amount is chosen close to the real business volume; ordinary enough not to attract attention.

Compromise of the supplier’s mailbox

This is the hardest scenario to detect. The attacker has got into the supplier’s real mailbox, read the correspondence for weeks and, during the payment period, sent a new IBAN as a reply in an existing email thread. The message comes from the real domain; it passes SPF, DKIM and DMARC; the correspondence history is real. An inbox rule is usually also added to the supplier’s mailbox to hide the replies. Technically, everything is clean.

Look-alike domains and display name spoofing

DMARC p=reject only prevents the exact use of your domain; details are in the moving DMARC to p=reject guide. That’s why attackers turn to these routes:

  • Look-alike domains: letter substitution (examp1e.com), adding or removing letters, a different top-level domain (example.com or example.co instead of example.com.tr), adding a hyphen (example-tr.com).
  • Internationalized domain name (IDN) look-alikes: characters from other alphabets that resemble Latin letters; the domain appears in DNS in Punycode form starting with xn--.
  • Display name spoofing: The address is [email protected], the display name is “General Manager’s Full Name”. It works because mobile clients usually show only the display name.
  • Reply-To redirection: The From address looks innocent, but Reply-To is set to another address; replies go to the attacker.

The open-source dnstwist tool can be used to monitor variations of your own domain. The command below generates possible permutations and lists only the registered ones:

dnstwist --registered example.com

Domains in the output with an MX record have priority. A look-alike domain that can receive email may be a sign of an attack in its preparation stage. Run the same scan for your critical suppliers’ domains too; the fake invoice will come in their name.

Technical controls

From an enterprise cybersecurity standpoint, technical controls don’t stop BEC on their own. Their job is to give the user a cue to be suspicious and to catch signs of compromise early. The examples below are for a Microsoft 365 environment.

External sender tagging. Exchange Online can add a built-in tag to external messages in supported Outlook clients:

Set-ExternalInOutlook -Enabled $true
Get-ExternalInOutlook

Impersonation protection. In Defender for Office 365 anti-phishing policies, executives and finance users can be added to the protected users list, and critical supplier domains to the protected domains list; details are in the Defender for Office 365 configuration guide.

Inbox rule monitoring. In a compromised mailbox, rules are the first place to look in the console. The typical trace is rules that delete, mark as read or move to a folder nobody opens, such as RSS, messages containing words like invoice, payment, IBAN or wire transfer:

Get-InboxRule -Mailbox [email protected] |
  Select-Object Name, Enabled, SubjectOrBodyContainsWords, MoveToFolder, DeleteMessage, ForwardTo, RedirectTo

Search-UnifiedAuditLog -StartDate (Get-Date).AddDays(-30) -EndDate (Get-Date) `
  -Operations New-InboxRule,Set-InboxRule,UpdateInboxRules -ResultSize 5000

Blocking automatic external forwarding. To stop the attacker from continuously copying correspondence, automatic forwarding can be turned off in the outbound spam policy; exceptions needed for business are defined with a separate policy:

Set-HostedOutboundSpamFilterPolicy -Identity Default -AutoForwardingMode Off

Identity protection. Mailbox compromise usually starts with stolen passwords or sessions. Phishing-resistant MFA and risky sign-in policies on finance and executive accounts come first; the topic is covered in the Entra ID Protection risk policies guide.

The payment change verification process

All the scenarios share the same gap: a request that came by email is verified by email. The control is to require verification through an independent channel. The process should be written, approved by finance management and applied to everyone without exception, including executives themselves. A process that allows exceptions isn’t a process.

Request type Verification method Approval
Supplier IBAN change A call back to the number registered on the supplier record Two people, change record
First payment to a new IBAN A second confirmation independent of the change request Finance manager
An urgent transfer request from an executive Voice confirmation from the executive’s known number The normal authorization matrix, no exceptions
New supplier registration Checking against tax details and official documents Purchasing and finance
Different account details on an invoice Comparison with the master contract and previous payments Accounting lead

The number used for the call back isn’t taken from the email, signature or invoice containing the change request; the attacker changes those too. In the ERP or accounting system, every change to the IBAN field should be recorded with who made it and on what evidence. Training for finance staff should specifically cover the “urgency and confidentiality request” pattern; for a program supported by realistic scenarios, see the cybersecurity training and phishing simulation service.

First steps during an incident

If the payment has gone out, the clock is ticking. The order below puts limiting the financial loss and preserving evidence first:

  1. Call the bank. Contact the sending bank immediately to have the transfer stopped or recalled; keep the receiving bank’s details and the transaction reference ready.
  2. Preserve the evidence. Don’t delete the related messages; export them with full headers. How to examine headers is explained in the suspicious email header analysis guide.
  3. Identify the source. Did the message come from a look-alike domain, or from inside the real supplier’s mailbox or your own? This distinction determines the next steps.
  4. Contain the compromised account. Reset the password, end the sessions (with Revoke-MgUserSignInSession -UserId), and review registered MFA methods, inbox rules, forwarding and app consents. Resetting the password alone isn’t enough; an MFA method or app consent added by the attacker stays in place.
  5. Widen the scope. Search for other messages from the same sender or a look-alike domain; identify outbound messages from the compromised account and notify the parties concerned.
  6. Make the notifications. Work with your legal department to file a criminal complaint with the police and prosecutor. If personal data was affected, assess the data controller’s obligation to notify the Personal Data Protection Board and the data subjects under Article 12 of Law No. 6698, together with the Board’s current decisions, from KVKK sources.

Who owns these steps should be written down before an incident. A tabletop exercise that brings finance, IT and legal to the same table quickly shows the real gaps in the process: where the call-back number is recorded, who will call the bank on a weekend, who will lock the mailbox. For setting up the technical controls on the email side together, the email security and phishing protection service can also be considered.

Frequently asked questions

Does DMARC p=reject stop BEC attacks?

Partly. DMARC stops the attacker from using your exact domain in the From field. Messages from a look-alike domain, a free mail account or a compromised real supplier mailbox pass DMARC. DMARC is necessary against BEC, but not enough.

If the supplier’s mailbox has been compromised, which technical control helps?

Because the message comes from the real domain with valid authentication, filters usually don’t catch it. The decisive control is the process: verifying a change in payment details by phone using previously recorded contact details, not those in the email, and requiring additional approval for the first payment to a new IBAN.

Is it better to add the external email tag to the subject line?

A prefix in the subject line is visible in every client, but it piles up in reply chains and, because it changes the message content, can affect signature verification in some cases. If you use Exchange Online, the built-in External tag in Outlook clients is cleaner; if there are clients that don’t support it, the two can be considered together.

What should be done first when we notice a payment has been made to a fake invoice?

The first step isn’t technical but financial: call the bank immediately and ask for the payment to be stopped or recalled. Then keep the related emails with their headers without deleting them, investigate signs of mailbox compromise and start the legal notification process.

Support

Contact us for support or a discovery session on this topic.

Contact form
[email protected]

Eretek Cybersecurity

Firewall, VPN, endpoint security and threat prevention. We reduce the attack surface and measure and report the risks.

Knowledge Center

Other articles