Chapter 17

Business Email Compromise Playbook

BEC is the costliest cybercrime category by dollar loss. Unlike ransomware, it leaves no encrypted files and no ransom note — only a completed wire transfer and a trail of inbox rules designed to hide the evidence. The window to recover funds is measured in hours.

Scenario

A CFO approves a $2.3 million wire transfer to a supplier based on instructions received via email from the CEO's M365 account. The email is real — it originated from the CEO's actual account, not a lookalike domain. The email thread looks legitimate: it references an ongoing acquisition deal discussed in prior emails, uses the CEO's writing style, and was followed up with a phone call from someone who sounded like the CEO (later determined to be AI-synthesized voice).

When the actual CEO calls the CFO the following day about the acquisition, the fraud is discovered. Finance reviews the wire and confirms it was processed and cleared into an overseas account 18 hours earlier. The wire recall window has likely closed.

Investigation reveals: the CEO's M365 account was compromised via phishing 11 days before the wire transfer. During those 11 days, the attacker read emails to understand ongoing business deals, created an inbox rule to forward copies of all finance-related emails to an external address, and monitored communications until they found the right moment to inject fraudulent wire instructions into an active business email thread.

Why BEC Works: The Account Access Advantage

Traditional email fraud used spoofed domains or lookalike email addresses. BEC evolved past this: attackers compromise the actual email account of a trusted executive or finance team member and conduct the fraud from the legitimate account. There is no spoofed domain to detect. The email passes all technical authentication checks (SPF, DKIM, DMARC) because it came from the real account.

  How BEC Progresses From Account Compromise to Wire Fraud
  ═══════════════════════════════════════════════════════════════════

  Phase 1: Account Compromise (T-11 days)
  ─────────────────────────────────────────────────────────────────
  Entry: phishing email with fake M365 login prompt.
  CEO enters credentials on attacker-controlled phishing page.
  Attacker uses credentials to authenticate to M365.
  MFA bypass: attacker uses Evilginx/AiTM proxy to steal session cookie,
  bypassing MFA entirely (password AND MFA token captured in real time).

  Phase 2: Environment Reconnaissance (T-11 to T-2 days)
  ─────────────────────────────────────────────────────────────────
  Attacker reads email — does not send anything yet.
  Goal: understand ongoing deals, payment processes, counterparties,
  CEO's writing style, legitimate payment amounts and recipients.
  Create inbox rule: auto-forward finance emails to attacker.
  Create inbox rule: delete or move to subfolder any email
  that might warn the CEO (IT security alerts, MFA prompts, etc.)

  Phase 3: Wire Fraud Execution (T-2 to T-1)
  ─────────────────────────────────────────────────────────────────
  Identify active business deal with legitimate payment context.
  Inject fraudulent wire instructions into existing thread
  (using CEO's account, from CEO's actual email address).
  Include detail from prior emails to appear legitimate.
  Optionally: AI voice synthesis to confirm "CEO approval" by phone.

  Phase 4: Funds Transit (T-1 to T+N days)
  ─────────────────────────────────────────────────────────────────
  Wire sent to first-hop account (often US bank).
  Within hours: funds transferred to second-hop international account.
  Within 24 hours: funds may be in cryptocurrency or third country.
  Recovery window closes as funds move through transfer chain.

Immediate Response: The Wire Recall Race

BEC response has a financial track and a security track running simultaneously. The financial track operates on a time-critical basis: wire recall has a narrow window. The security track must understand the account compromise scope. Both must happen in the first two hours.

  BEC Response: First Two Hours
  ═══════════════════════════════════════════════════════════════════

  FINANCIAL TRACK (starts immediately)
  ─────────────────────────────────────────────────────────────────
  T+0:00  Confirm the wire was fraudulent (not a legitimate payment)
  T+0:05  Call the originating bank (the bank the wire came from)
          Request: wire recall / recall of funds
          Provide: wire reference number, beneficiary name, amount, date
          Be explicit: "This was a fraudulent wire; I am requesting recall."
  T+0:10  Call the receiving bank (if identifiable from wire instructions)
          Request them to freeze the destination account
          Provide: your bank's contact name + recall request reference
  T+0:15  File FBI IC3 complaint at ic3.gov
          The FBI's Financial Crimes Unit has a direct relationship with
          major banks and can issue freeze requests faster than most
          organizations can through normal channels
  T+0:20  Notify your bank's fraud team in writing (email)
          Document: everything about the transaction in writing NOW

  Success rate: Wire recall drops dramatically after 24 hours.
  After 72 hours: odds of recovery are below 15%.
  After funds reach overseas accounts: recovery through civil channels only.

  SECURITY TRACK (starts in parallel)
  ─────────────────────────────────────────────────────────────────
  T+0:00  Reset the compromised account credentials immediately
          Enable/verify MFA is active
          Revoke all active sessions (M365: Sign out all sessions)
  T+0:15  Search for inbox rules created by attacker (see below)
  T+0:30  Pull audit logs: what did the attacker access? What emails?
  T+1:00  Determine scope: which other accounts may be compromised?
  T+1:30  Notify potentially affected counterparties (suppliers, partners
          who were in email threads with the compromised account)

M365 Audit Log Forensics

M365 unified audit logs are the primary forensic source for BEC investigations. They record every authentication event, email access event, inbox rule modification, and email send event. The key events to review are in the Exchange audit log and the Azure AD sign-in log.

Audit eventOperation nameWhat it tells you
Email accessed by attacker (without downloading) MailItemsAccessed Which emails the attacker read. Requires M365 E5 or Purview Audit Premium. Shows: items read, timestamp, client IP, authentication type
Inbox rule created Set-InboxRule / New-InboxRule What rules were created, when, and from what IP. Rule details include: conditions (forward if subject contains "invoice"), actions (forward to, delete, move to folder)
Email sent from account Send All emails sent from the compromised account — including fraudulent wire instructions. Shows recipient, subject, timestamp, sending IP
Successful authentication UserLoggedIn (Azure AD) Every authentication to M365 from the account. Shows: IP, location, device, authentication method, whether MFA was used or bypassed
OAuth app consent Consent to application If attacker consented to a malicious OAuth app for persistent access — survives password reset
Email forwarding rule Set-Mailbox with ForwardingAddress Mailbox-level forwarding (different from inbox rule forwarding) — more persistent, set at mailbox level not client
Why MailItemsAccessed is a critical BEC forensic signal

MailItemsAccessed (introduced in M365 as part of Purview Audit Premium, formerly Advanced Audit) records every time email items are accessed, not just sent or received. This means you can reconstruct exactly which emails the attacker read during their 11-day reconnaissance phase. This is forensically valuable for two reasons: it defines the breach scope (which sensitive business information did the attacker have access to), and it identifies which business counterparties need to be notified about potentially compromised communications. Without MailItemsAccessed, you can only know what the attacker sent and received — not what they read. The difference matters enormously for M&A deals, legal matters, or strategic plans that were discussed in emails during the dwell period.

  Finding and Removing Attacker Inbox Rules in M365
  ═══════════════════════════════════════════════════════════════════

  Via M365 Admin Center:
  Exchange Admin Center → Recipients → Mailboxes
  → [Select affected mailbox] → Manage mail flow settings
  → Review inbox rules for any rules not created by the user

  Via PowerShell (audit all inbox rules for a user):
  Connect-ExchangeOnline -UserPrincipalName admin@contoso.com
  Get-InboxRule -Mailbox "ceo@contoso.com" |
    Format-List Name, Description, Enabled, ForwardTo,
    DeleteMessage, MoveToFolder, ConditionIsString

  What to look for:
  → Rules that forward to external addresses
  → Rules that delete emails matching keywords (security, alert, MFA,
    unusual sign-in, password reset)
  → Rules that move emails to obscure subfolders
  → Rules created from IPs not associated with the user
  → Rules with creation timestamps during the compromised period

  Important: removing inbox rules does not end the attacker's access
  if they have active sessions or OAuth app permissions. Session
  revocation and password reset FIRST, then rule removal.

Scoping the BEC Compromise

BEC investigations frequently reveal broader compromise than the single fraudulent wire that triggered the incident. A thorough scope assessment asks: how long was the account compromised, what was accessed, were other accounts targeted, and what financial controls were exploited?

Scope questionWhere to find the answer
When was the account first compromised? Azure AD sign-in logs — find first authentication from unusual IP or location. Correlation: first inbox rule creation timestamp.
What emails did the attacker read? MailItemsAccessed audit events (requires E5/Purview Audit Premium)
Were other accounts compromised? Check if attacker used compromised account to send phishing to internal contacts. Check Azure AD for unusual auth from other accounts at similar times.
What financial controls were bypassed? Review the fraudulent wire approval chain — was dual authorization required? Was the beneficiary a new account? What controls would have caught this?
What business information was exposed? MailItemsAccessed + review of email contents from the dwell period — identify: M&A discussions, customer data, legal privileged communication, trade secrets
Are other wires or payments at risk? Review all pending payments submitted via email during the dwell period — any payment instruction received via email may be fraudulent

The Financial Controls That BEC Exploits

BEC succeeds because it exploits gaps in financial control processes designed for a world without email-based business communication. Understanding which controls failed — and which controls would have prevented it — is essential for both the PIR and for reducing recurrence risk.

  Financial Controls and BEC Bypass Methods
  ═══════════════════════════════════════════════════════════════════

  CONTROL: Dual authorization for wires above threshold
  BEC BYPASS: Attacker compromised one authorizer and socially
  engineered the second (or the organization had single-auth for
  "urgent" CEO-directed transfers)
  FIX: No exceptions to dual authorization. "Urgency" is a social
  engineering flag, not a justification for exception.

  CONTROL: Callback verification for wire changes
  BEC BYPASS: Callback made to attacker's AI-synthesized voice clone,
  or to phone number supplied by the attacker in the email
  FIX: Callback must use a pre-established, verified phone number
  from the company directory — never the number in the email

  CONTROL: New beneficiary review process
  BEC BYPASS: The "supplier" account was different from the real supplier's
  account on file, but the review was cursory or bypassed because the
  wire was "CEO-directed"
  FIX: Any new beneficiary account requires independent verification via
  a phone call to the supplier using their established contact information

  CONTROL: Finance team suspicion of unusual wire amounts
  BEC BYPASS: $2.3M was within the range of real transactions for this
  company; the attacker chose an amount that wouldn't trigger "that seems
  too large"
  FIX: Verify the specific transaction's context with the apparent requester
  via a separate, pre-established channel — not a reply to the email thread

  CONTROL: Business email authentication (DMARC/DKIM/SPF)
  BEC BYPASS: The email DID pass all authentication — it came from the
  legitimate, compromised account
  FIX: Technical email authentication does not prevent account compromise BEC.
  The countermeasure is account security (MFA, phishing-resistant auth), not
  email authentication.

Q & A

Q: The attacker was in the CEO's M365 account for 11 days. What data exposure notifications are likely required?

Notification obligations depend on what data was in the accessed emails. If the CEO's email contained: (1) Personal data of employees (compensation, HR matters, health information) → potential state privacy law notification requirements. (2) Customer PII → GDPR notification if EU customers are involved (72-hour window to supervisory authority), state breach notification laws. (3) Legal privileged communications → notify legal counsel immediately; there may be privilege waiver implications. (4) M&A or strategic non-public information → legal and SEC counsel (public companies may have disclosure obligations). The scope analysis from MailItemsAccessed logs directly drives the notification decision. Organizations without MailItemsAccessed (E3 licenses only) face a harder choice: they may need to assume all email in the account during the dwell period was accessed and notify accordingly.

Q: MFA was enabled on the CEO's account. How did the attacker bypass it?

Adversary-in-the-Middle (AiTM) phishing is the primary method. The attack works like this: the attacker sets up a reverse proxy (using tools like Evilginx, Modlishka, or Microsoft-developed-for-research tools now commercialized in criminal toolkits). The phishing link sends the CEO to the attacker's proxy server, which forwards the request to the real M365 login page. The CEO sees a legitimate-looking M365 login, enters their username and password (which the proxy captures in real time), then the real M365 prompts for MFA. The CEO completes the MFA challenge (proxy forwards this too), and M365 issues a session cookie to the proxy. The proxy now has a valid, authenticated session cookie that includes the MFA-satisfied state. The attacker uses this cookie to authenticate as the CEO without needing the password or MFA. The defense: phishing-resistant MFA (FIDO2 hardware keys, passkeys) that binds the MFA assertion to the legitimate origin domain. SMS OTP and authenticator app push notifications are both vulnerable to AiTM; FIDO2 is not.

Q: Twenty-four hours have passed and the wire recovery attempt failed. The $2.3M is gone. What happens next?

Recovery shifts from urgent financial response to legal and insurance track: (1) Maintain and preserve all evidence — M365 audit logs, wire transfer records, all related emails, bank communications. These support the insurance claim and any potential prosecution. (2) File FBI IC3 complaint if not already done. The FBI has successfully recovered BEC funds even months after the wire through civil forfeiture and international cooperation, though success rate drops significantly over time. (3) Engage cyber insurance — most policies cover BEC financial fraud losses but require timely notification (check your policy's reporting window). The claim will require the full incident documentation. (4) Legal counsel review — assess civil recovery options against the receiving bank or correspondent banks, and evaluate whether any third parties (payment processor, bank that failed to flag suspicious activity) have liability. (5) PIR and financial controls improvement — implement the controls that would have prevented this before the next attempt.

Q: The CFO is asking: "How do we make sure this never happens again?" What are the top three controls?

For preventing account compromise: deploy phishing-resistant MFA (FIDO2 hardware tokens or passkeys) for all executives and finance staff. Standard push-notification MFA is bypassed by AiTM phishing. For preventing wire fraud: implement an out-of-band verification process for all wire transfers above a threshold — a callback to a pre-established phone number from the company directory (not from the email) that verifies the transfer with the requester directly. This control is simple and inexpensive, and it catches both account compromise BEC and impersonation BEC. For detecting account compromise: enable MailItemsAccessed in M365 Purview Audit Premium and configure alerts for unusual sign-in locations, new inbox rules, and external forwarding. An inbox rule forwarding to an external Gmail should generate an immediate alert to IT security — it almost never has a legitimate explanation. These three controls together address the account security, the payment process, and the detection gaps that allowed this incident to proceed for 11 days undetected.

Q: How do you explain BEC to executives who think "we have email security" and are confused that the CEO's real account was used?

Frame it in terms they understand: "Our email security protects against someone impersonating our CEO by faking their email address. This attack was different — the attacker had the actual keys to the CEO's email account, the same way a thief with your actual house key doesn't need to break a window. Once they had the account, everything the CEO sends looks completely legitimate to all our email security systems, because it is technically coming from the CEO's real account. The countermeasure is not better email security — it's account security (making it much harder to steal the key), and financial process security (requiring verification through a second channel before sending large wires, regardless of how legitimate the email looks)." Executives understand account access and key management as a concept; the analogy transfers the risk understanding without requiring technical knowledge of M365 authentication.