ADAuditcover Direct
Guides › Security Controls

Seven Payroll Diversion Fraud Mistakes HR and Finance Teams Make, and the Fixes

Payroll diversion fraud slips past the callback controls you built for vendor payments, and the loss often lands under your deductible. Here are the gaps attackers use and the fix for each.

Auditcover Direct editorial team · November 2026 · 8 min read

Leaving payroll outside the controls you built for vendor payments

Most finance teams have put real work into vendor-payment controls: callbacks to a known number before bank details change, dual approval on new payees, alerts on unusual transfers. Then they leave the employee master file wide open. That gap is what payroll diversion fraud exploits. The attacker doesn't invoice anyone. They pose as an existing employee, change where that person's net pay goes, and collect one or two pay cycles before anyone notices.

It works because payroll changes feel routine and low-value. A single redirected salary might be USD 2,000 to 12,000, so nobody treats it like a six-figure supplier change. Attackers rarely stop at one employee, though, and the controls that would catch them are the ones you already run in accounts payable.

Advertisement

What to do instead: Put employee bank-detail changes in the same risk category as vendor changes, and write that into your payment-controls policy. That means one owner (usually the controller or head of payroll), one verification standard and one audit trail for both. When your auditors or insurer ask about change-of-bank-details controls, your answer should cover employees as well as suppliers.

Accepting direct deposit changes by email alone

The classic version of direct deposit change fraud is a short email to HR or payroll: "I've switched banks, please update my account before Friday's run." It comes from a lookalike domain, a personal webmail address the employee supposedly uses when traveling, or a compromised genuine mailbox. The last is the hardest to spot, because nothing about the sender looks wrong.

Email proves nothing about identity, and replying to the sender just confirms the request with the attacker.

What to do instead:

  • Refuse bank changes submitted by email, chat or text. Point employees to one authenticated channel, normally the HR or payroll self-service portal.
  • For any change that arrives outside that channel, call back on a number taken from the HR record. Never use a number from the request itself, or one added recently.
  • Define red-flag triggers that call for extra checks: urgency tied to the next pay run, a new bank or e-money account in another country, a request made while the employee is on leave, or several changes in one week.

Train payroll staff to treat a polite, urgent request from a senior executive as more suspicious, not less. Executives get impersonated precisely because people hesitate to question them.

Relying on passwords or SMS codes to protect HR self-service

Moving changes into a self-service portal only helps if the portal is hard to break into. Many HR and payroll platforms still run on password-only logins, or on SMS codes and push approvals that phishing kits and MFA-fatigue attacks can get around. An HR portal account takeover often follows a predictable pattern: the attacker phishes the employee's credentials, signs in, swaps the registered MFA method or phone number, then changes the bank details, all in one session and frequently outside business hours.

What to do instead:

  • Put the HR portal behind your corporate single sign-on so it inherits your identity controls, and require phishing-resistant MFA (FIDO2 security keys or passkeys) for any session that changes bank details or contact information.
  • Use conditional access to block or step up sign-ins from new devices, anonymizing networks or countries where the employee doesn't work.
  • Treat a change of MFA method, phone number or email followed shortly by a bank change as a high-severity alert that holds the payment.

The cost is modest. Hardware keys typically run USD 25 to 75 each in 2026, and passkeys on managed devices cost little beyond setup time. Prices vary by vendor and volume, so confirm with suppliers. If you can't roll this out to everyone, start with payroll and HR administrators and your highest earners.

Letting a change pay out in the very next run

When a bank change made on Wednesday pays out on Friday, the attacker has the money before the real employee notices anything. Speed makes this fraud profitable, so slowing it down is one of the cheapest controls available.

What to do instead:

  • Notify on both channels. Alert the employee's previous email and phone as well as the new ones, so the real person sees the change even if the attacker controls the new contact details.
  • Set a cooling-off rule. For example, changes made within five business days of a pay date take effect the following cycle, or the first payment to a new account goes out only after the employee confirms through a separate channel.
  • Use account-name checks where your country offers them. The UK's Confirmation of Payee, the EU's Verification of Payee for euro transfers, and penny-drop account validation (used in India and offered by many payment providers elsewhere) can flag a mismatch between the account holder's name and the employee. Where there's no national scheme, ask your bank or payroll provider what validation tools they offer. In the US, an ACH prenote confirms the account is valid but not who owns it.

A name mismatch shouldn't block the change on its own, since joint accounts and name variations are common. It should route the request to a manual callback.

Letting one person change and release pay, or reset access without checks

In smaller teams, one payroll administrator often edits employee records, runs payroll and approves the file. Attackers who target that person, or who social-engineer the IT help desk into resetting an employee's password and MFA, meet no second line of defense. Help desk resets are a common entry point for employee impersonation fraud. The caller can know the employee's name, manager and start date, all easy to pull from professional networking profiles.

What to do instead: Separate who makes a bank change from who approves the payroll file. Run an exception report each cycle listing every bank change since the last run, reviewed by someone outside payroll. For help desk resets, require identity proofing stronger than knowledge questions: a video check against the HR photo, manager confirmation, or a code sent to a device already enrolled.

If you outsource payroll to a global provider or an employer-of-record platform, don't assume they handle this. Ask how they verify bank changes, whether they hold first payments to new accounts, and who carries the loss when their process fails. Contracts are often silent on that last point.

Assuming your cyber policy will pay for diverted wages

Many buyers get caught out here. You still owe the employee their wages, so the diverted amount is a direct loss to the company. Whether insurance pays depends on how the fraud happened and how your policy is worded, and payroll fraud insurance coverage often turns on small definitions.

The table shows how these losses are commonly treated. Wording varies widely by insurer and country, so use it as a starting point for a conversation with your broker.

How the fraud happenedCoverage section that may respondCommon sticking points
Payroll staff tricked by an email or call into changing detailsSocial engineering or fraudulent instruction cover (crime policy or cyber endorsement)Sublimits are often modest relative to the policy limit; some wordings only cover vendor or client impersonation; many require a documented callback as a condition of cover
Attacker logged into the HR portal and changed details directlyComputer fraud or funds transfer fraud (crime); cyber crime sectionsInsurers may argue the entry was "authorized" because valid credentials were used; disputes over whether the loss was direct
An insider changed details to their own accountEmployee theft (crime policy)Usually excluded from cyber policies; discovery and notice deadlines apply
HR records exposed during the takeoverCyber first-party breach responseCovers investigation and notification costs, not the diverted wages

Then there's the retention. Say a 250-person company (hypothetical) loses USD 18,000 across three diverted salaries and carries a USD 25,000 retention. Nothing is recoverable. For most mid-size buyers, controls will save more money than insurance on this type of fraud. Ask a licensed broker whether your fraudulent instruction wording names employee impersonation, and record the callback steps you actually follow.

Treating "I didn't get paid" as a payroll error

The first sign is usually an employee saying their salary hasn't arrived. Teams often spend days checking bank cut-offs and reissuing payments before anyone considers fraud. By then the money has usually been moved on.

What to do instead: Write a short playbook that payroll can start without waiting for security.

  1. Within the first hour, check the employee's change history. If the details changed recently, treat it as fraud, ask your bank to recall the payment and alert the receiving bank.
  2. Freeze further changes on that record, reset the employee's credentials, and check every other record changed in the past 60 days for the same pattern.
  3. Notify your insurer or broker straight away. Late notice is a common reason for disputed claims.
  4. Keep portal sign-in logs, email headers and help desk tickets, and report the fraud to your national fraud or cybercrime reporting body.
  5. Assess whether personal data was exposed. Regimes such as the GDPR can require regulator notification within tight deadlines, and India, Brazil, the Gulf states and others have their own breach-notification rules. Take advice from qualified counsel.

Before your next pay run, pull every employee bank change from the last quarter and check whether each was verified through a channel the attacker couldn't control. If you can't show that for every change, fix verification before you spend anything on higher insurance limits.

Frequently asked questions

Is payroll diversion fraud covered under a standard cyber insurance policy?

Sometimes, but not reliably. Many cyber policies cover social engineering only through an optional endorsement with a low sublimit, and some define it narrowly as impersonation of vendors or clients. Ask your broker whether the wording names employee impersonation and whether a callback is a condition of cover.

What is the single most effective control against direct deposit change fraud?

Out-of-band verification using contact details that were on file before the request arrived, combined with phishing-resistant MFA on the HR portal. Together they stop both the email-based version and the account-takeover version. A short hold on first payments to a new account adds a useful safety net.

Do we have to pay the employee again if their wages were diverted?

In many countries, yes. The employer's duty to pay wages is generally not discharged by paying a fraudster, which is why the diverted amount usually becomes the company's loss. Employment rules differ by jurisdiction, so confirm your obligations with local employment counsel.

Our payroll is outsourced. Is the provider responsible for losses?

It depends on your contract and on whose process failed. Many provider agreements limit liability or don't address fraudulent bank changes at all. Review the service agreement, ask how the provider verifies changes, and negotiate clear allocation of fraud losses at renewal.

Written by the Auditcover Direct editorial team. Prices, features and rules change often - confirm details with the vendor or a licensed professional before you decide.