Anonymized incident investigation | Supplier payment workflows
The fraud came from a real business conversation
An attacker posed as a prospective buyer, built an ordinary sales conversation and used a shared-document lure to obtain a Microsoft 365 sign-in and a live MFA code. Persistent mailbox access then supported fraudulent payment instructions to customers. A customer question exposed the problem, and the user’s password change ended the observed access. ACS Cyber subsequently reconstructed the activity, found a second fraud path and removed a hidden inbox rule with authorization.
- Microsoft 365
- Business email compromise
- Payment-change verification
- No financial details published
If a customer may have sent money to fraudulent instructions, contact the bank through its known fraud channel immediately, preserve the payment and message evidence, and involve the appropriate insurer, counsel and law-enforcement contacts. Do not wait for a routine IT consultation. No recovery of funds is claimed or promised in this story.
Evidence standard
A documented investigation with client identifiers removed
This account is based on a confidential incident report supplied for anonymized publication. Company and personal names, customer and supplier identities, domains, bank details, exact dates, amounts and account fingerprints are intentionally omitted.
The review used Microsoft 365 mailbox and sign-in records, directory and device changes, connected-application information, message trace and authorized content-search evidence. The affected user’s computer was outside the review, so the report does not claim that every endpoint was checked or that all remediation was complete.
Who did what
- The user paused a payment conversation and changed the account password
- Business leaders warned affected contacts and disabled an unauthorized device record
- ACS Cyber began with a read-only review and preserved relevant evidence
- ACS Cyber removed the malicious rule with approval and rechecked available records
Telltale signs
The early clues looked like normal business
A patient fake-buyer conversation
The attacker did not begin with an obvious invoice demand. A normal inquiry developed into a document-sharing exchange, making the later sign-in request feel connected to real work.
A document lure with an access-code prompt
The lure asked the user to copy a code and open the shared material. A familiar file-sharing brand or a real email thread does not validate a sign-in destination.
An unfamiliar computer joined the account
A new device registration shortly after the sign-in was a high-value identity signal. Device-join permissions needed the same scrutiny as the password.
Supplier replies disappeared
A hidden inbox rule moved genuine replies out of view. A customer then questioned unexpected banking instructions and a legitimate supplier reported that a payment had not arrived.
Inside the mailbox
How ordinary records revealed the fraud path
01
Automated reading continued after the initial sign-in
Audit evidence showed repeated access to email and attachments related to real orders, invoices and payments. This gave the attacker the context to make a fraudulent request sound plausible.
02
Real replies were hidden and fraudulent messages were purged
The mailbox rule concealed supplier and customer replies. Deleting the attacker’s sent message did not eliminate all evidence: its content remained in a customer’s quoted reply.
03
A second fraud path used a look-alike supplier identity
A company-wide search connected altered invoice material and a look-alike domain to the same fraud activity. The scope was broader than the one conversation that first raised concern.
04
Available tenant records were checked beyond the mailbox
The team reviewed account sign-ins, mailbox-rule and permission changes, directory activity, registered devices and application access. The evidence tied the observed attacker activity to the affected account, not every mailbox in the tenant.
05
Containment was verified only within the available evidence
The password change ended the observed unauthorized access, and subsequent record checks did not find new attacker activity. That is a bounded containment finding, not a guarantee about unexamined devices or future attempts.
What reduced the risk
Immediate containment and longer-term control are different jobs
| Stage | Action documented in this incident | What it did—and did not establish |
|---|---|---|
| User response | Paused the suspicious payment conversation, changed the password and warned contacts. | Ended the observed access; did not remove the hidden rule or reverse payments. |
| Business response | Disabled the unauthorized device record and started account cleanup. | Reduced an identity path; broader device and account configuration still needed review. |
| ACS investigation | Reconstructed activity, preserved records, found the additional fraud path and removed the hidden rule with approval. | Restored visibility of hidden mail and established a clearer scope; did not prove recovery of money or data. |
| Recommended follow-up | Phishing-resistant authentication, controlled device enrollment, managed endpoints, application review, email protection and monitored alerts. | Risk-reduction plan matched to observed weaknesses; implementation and ongoing ownership must be explicit. |
The bank and the affected parties own payment-recall decisions. Counsel and the insurer own legal and coverage decisions. ACS Cyber supplies technical investigation and agreed security work within that wider response.
Process plus technology
A safe payment change needs an independent check
Email controls alone cannot decide whether a payment change is legitimate. Staff should pause when bank details change, use a known phone number from an independent record, involve a second authorized person and retain a written approval record. A reply to the existing email thread is not independent verification.
ACS Cyber can help connect that business process to identity, email, endpoint and monitoring controls through Business Email Compromise response and managed detection and response. For the surrounding IT operations, the Atlantic Computer Systems security team provides the parent-company technology path.
Stop-and-verify routine
- Pause on changed bank details or unusual urgency
- Call a known contact using independently stored details
- Require the organization’s approved second-person review
- Do not reply to the suspect thread to validate it
- Preserve messages and payment records if fraud is suspected
Related evidence and guidance
Learn from the response without assuming the same outcome
Google Workspace account takeover
A separate anonymized incident involving a trusted-sender lure, session revocation and browser evidence.
Signs of compromise
Review identity, mailbox, endpoint and payment warning signs that deserve a structured response.
The first hour of Business Email Compromise
Use a call-first sequence for the bank, identity containment, evidence and responsible escalation.
Law-firm security support
ACS Legal IT and ACS Cyber help legal organizations connect daily IT and security responsibilities. This was not a law-firm incident.
Make the next suspicious request easier to handle
Give unusual account and payment activity a clear review path
If money may have moved, call the bank first through a known fraud channel. For an active account compromise, call our incident line at 1-650-300-7557 (after-hours emergency support is available). For a planned review of identity, email, endpoints and payment-change controls, book a security consultation.