Cyberpediaby Brain

The hacked supplier: vendor email compromise

An attacker takes over a real supplier's mailbox and writes to its customers to change bank details. Here is how it works.

On this page
  1. 1In 30 seconds
  2. 2What it is
  3. 3Straight answers
  4. 4Inside the attacker's head
  5. 5What you see on screen
  6. 6Under the hood
  7. 7Case files
  8. 8Sources
  9. 9Cite this article

In 30 seconds

Where it happens
On an ongoing invoicing thread with a known supplier. In one case described by Proofpoint, two people working in finance and accounting were targeted[2].
What the attacker wants
To get a legitimate invoice paid into a bank account the attacker controls[3].
The reflex
Before any change of bank details, call the supplier on a number taken from a known contact list, never from the message received[1].
If it has already happened
According to CISA, when the bank contacts the company to check a transfer, staff should delay the operation until further checks are made[1]. Proofpoint notes that these frauds rarely end with the funds being recovered[2].
Article map

What it is

A regular supplier writes to say that its bank details have changed. The message comes from its real address, inside the thread of an invoice in progress. Yet the supplier is no longer the one writing: an attacker has taken control of its mailbox and uses it to address its customers.

In June 2015, CISA (United States) described incidents in which a supplier's mailbox is compromised in order to change the bank account associated with that supplier[1].

Straight answers

What is a vendor compromise?

CISA defines business email compromise (BEC) as a payment fraud that works through the compromise of legitimate business email accounts, in order to obtain an unauthorized transfer[1]. In the variant discussed here, the compromised account is a supplier's[1]. MITRE ATT&CK notes that an attacker can use a compromised email account to hijack existing conversations with the targets they are interested in[4].

Why does the message get past the filters?

It comes from a legitimate account. Proofpoint explains that fraudulent messages sent from such an account pass email authentication checks (DKIM, SPF, DMARC)[2]. In a campaign described by Microsoft, a few of the targeted organizations had rules that automatically allowed email from trusted vendors, which let the attacker through[5]. MITRE ATT&CK adds that attackers can target known accounts or domains, from which spam or phishing may escape filtering rules based on reputation[4].

Can you recognize the message with the naked eye?

Rarely. Proofpoint considers that conversation hijacking attacks are very difficult, even impossible, for users to identify[2]. In the case Proofpoint describes, the messages contained neither a link nor an attachment[2]. According to Proofpoint, these attacks can be among the most convincing a user will encounter[3].

What should you do before changing bank details?

CISA lists several methods[1]:

  • verify any change of payment instructions by calling the supplier; the number should not come from the message, but from a known contact list;
  • keep a record of the contact details of the supplier's people authorized to approve these changes, preferably in non-electronic form;
  • limit the number of employees who can approve or carry out transfers;
  • require dual approval for any transfer involving, among other things, new account numbers for existing partners, a new partner, or a country outside usual business patterns.

Proofpoint adds that its awareness training can enable users to report these messages as suspicious[2].

Inside the attacker's head

The attacker starts by gaining access to the account. MITRE ATT&CK lists several routes: collecting credentials through phishing, buying them on third-party sites, guessing them by brute force, or paying employees, suppliers or partners for access[4].

Next comes observation. CISA describes a reconnaissance phase during which attackers read the company's legitimate exchanges, sometimes by automatically forwarding received emails to an account of their own, until they feel comfortable enough to send their transfer instructions[1]. In the campaign Microsoft analyzed in 2022, the attacker checked financial emails and attachments every few hours and looked for threads where a payment fraud was feasible[6].

Once a transaction is spotted, the attacker replies within the existing conversation. The message is part of a thread the victim believes is authentic, which makes it more credible; Proofpoint cites requests to change banks justified by an audit or by COVID-19[2].

The attacker also covers their tracks. In the case Microsoft describes, the attacker created an inbox rule that moved incoming messages from the target's domain to the archive, marking them as read, then deleted their own replies[6]. The owner of the hacked account therefore does not see the target's replies[6].

Finally, some attackers change identity along the way. Proofpoint describes a "pivot" in which the attacker puts an account that looks like the supplier's in the "reply-to" or "cc" field. This allows the conversation to continue if the compromised account is cleaned up, and it complicates the investigation, because the logs of the supplier's mail gateway are lost[2].

The amounts explain the appeal: according to Proofpoint, attacks that exploit supplier relationships can cost from tens of thousands to several million dollars[3].

What you see on screen

An email from the supplier, from its usual address, in a thread about invoices that are really in progress[2]. The text asks to change the bank account used for payment and combines authority and urgency, two classic levers of these attacks[2].

Excerpt
Illustration (paraphrased message, neutralized addresses)
From: jane.doe@supplier[.]invalid  (the supplier's real address, hacked mailbox)
Subject: RE: [subject of the invoice in progress]
[Request to pay the next invoice into a new bank account,
 with a justification and an urgent tone]

A second message may follow, from an almost identical address. Proofpoint gives supp1ier.com instead of supplier.com as an example[2]. In the case it describes, this second email reused the same invoice numbers as the first and went to two targets, and the two messages reinforced each other[2].

Excerpt
Illustration (neutralized addresses)
Real domain:         supplier[.]invalid
Look-alike domain:   supp1ier[.]invalid

Under the hoodIT teams

The email is legitimate in the technical sense, since it leaves from the real account[2]. MITRE ATT&CK indicates that this technique cannot easily be prevented with preventive controls, because it relies on actions taken outside the company's defensive perimeter, and that much of the activity takes place out of sight of the targeted organization[4]. Hence the value of payment procedures, which act after the message rather than before it.

To take over the account, one method documented by Microsoft is adversary-in-the-middle (AiTM) phishing. The attacker places an intermediary server between the user and the real site, which lets them steal the password and the session cookie[6]. Microsoft specifies that this is not a flaw in multifactor authentication: with the cookie, the attacker is authenticated whatever the sign-in method[6]. In the 2023 campaign, the attacker then added a multifactor authentication method (a code received by SMS) to sign back in without being detected[5]. For Microsoft, resetting the password is then not enough: session cookies also need to be revoked and changes to multifactor authentication need to be rolled back[5].

Case filesReal cases

June 2015: CISA alert TA15-175A

On June 24, 2015, CISA published an alert on the rise in wire transfer fraud against American companies[1]. Funds go mostly to Asia[1]. Most incidents involve the mailbox of the chief executive or financial officer, others that of a supplier[1].

May 2021: more than 100,000 dollars targeted by a fake wine invoice

In May 2021, Proofpoint described an attempt to steal more than 100,000 dollars. The attacker posed as a wine seller, picked up an existing thread and asked the buying company to change its payment bank details, insisting that all communication go through email[3]. Faced with the lack of response, the attacker became firmer, then attached an invoice with a seal and stamp[3]. Proofpoint suspects a compromise outside its visibility, followed by impersonation through display name and reply-to address[3].

July 2022: a campaign that targeted more than 10,000 organizations

In an article of July 12, 2022, Microsoft describes an AiTM phishing campaign that, according to its data, attempted to target more than 10,000 organizations since September 2021[6]. After the session theft, the delay before the payment fraud could be as short as five minutes[6]. On one occasion, the attacker ran several attempts at the same time from the same mailbox, adding each new target domain to their rule[6].

June 2023: a campaign that started from a trusted supplier

On June 8, 2023, Microsoft described an attack against banking and financial services organizations, started from a compromised trusted supplier, then turned into a series of AiTM attacks and BEC activity across several organizations[5]. The first email came from a supplier trusted by the target, with a seven-digit code in the subject and a link to view a fax[5]. With the stolen mailbox, the attacker sent more than 16,000 emails to the victim's contacts, chosen from their recent exchanges[5]. The attacker read the replies of those who doubted the email's authenticity and answered them, probably to wrongly confirm that it was legitimate[5].

Sources

  1. CISA, “Business Email Compromise Continues to Swindle and Defraud U.S. Businesses”, 24 June 2015. https://www.cisa.gov/news-events/alerts/2015/06/24/business-email-compromise-continues-swindle-and-defraud-us-businesses (accessed 14 November 2026)
  2. Proofpoint, “Understanding BEC Scams: Supplier Invoicing Fraud”, 7 December 2020. https://www.proofpoint.com/us/blog/cybersecurity-essentials/understanding-bec-scams-supplier-invoicing-fraud (accessed 14 November 2026)
  3. Proofpoint, “BEC Taxonomy: Invoice Fraud”, 6 May 2021. https://www.proofpoint.com/us/blog/threat-insight/bec-taxonomy-invoice-fraud (accessed 14 November 2026)
  4. MITRE ATT&CK, “Compromise Accounts: Email Accounts, Sub-technique T1586.002”. https://attack.mitre.org/techniques/T1586/002/ (accessed 14 November 2026)
  5. Microsoft, “Detecting and mitigating a multi-stage AiTM phishing and BEC campaign”, 8 June 2023. https://www.microsoft.com/en-us/security/blog/2023/06/08/detecting-and-mitigating-a-multi-stage-aitm-phishing-and-bec-campaign/ (accessed 14 November 2026)
  6. Microsoft, “From cookie theft to BEC: Attackers use AiTM phishing sites as entry point to further financial fraud”, 12 July 2022. https://www.microsoft.com/en-us/security/blog/2022/07/12/from-cookie-theft-to-bec-attackers-use-aitm-phishing-sites-as-entry-point-to-further-financial-fraud/ (accessed 14 November 2026)