When a rep's mailbox is compromised, ask IT to cut off access, reset credentials in the right place, and investigate the account's reach.
Key takeaways
- Stop the attacker's access before anything else, because the mailbox may still be open while you read the thread.
- Reset the password where your identity system keeps it, since a change made only in the cloud may leave the stolen password working.
- Reach clients only by a route you already know works, because the rep's mailbox is the account the attacker controls.
- Check for inbox rules and forwarding settings the rep did not create, and ask IT whether any remain after the reset.
- Write down each action with its time, and keep your legal or compliance contact informed when the response begins.
When a prospect tells you your rep sent them something the rep never wrote, the first instinct is to read the message closely and figure out what happened. Resist that for a few minutes. The message is evidence, but the mailbox is probably still open. An attacker who can send mail as your rep can keep doing it while you read the thread. The first hour is about stopping access, not analyzing it.
This post covers the three things to do first, in order. It uses one example case so each step stays concrete. It is written for the revenue leader who has to get the right people moving quickly and who needs to know what to ask for. It does not replace your IT or security team, who will make the technical calls.
The example case
Imagine a senior account executive on a normal working morning. A procurement contact replies to a message the rep never sent. The message asks for a change to the bank details on a pending invoice. The rep is in a client meeting and has not seen it. The contact forwards the reply to you, and you check the rep's sent folder. There are several outgoing messages the rep does not remember writing. They went to people who are active in the rep's pipeline, and at least one of them refers to a deal the rep was negotiating last week.
That is a bad morning, but it is a common one. Several things are probably true at once. The attacker has a foothold in the rep's account. They may be reading the rep's email to learn who is involved in each deal. They may have set something up to hide their activity from the rep. And they can send messages that look exactly like the rep's, which is why the procurement contact believed the invoice request.
The scope of the problem is larger than the inbox. Microsoft's guidance on compromised email accounts describes what an attacker can reach: "Once an attacker gains access to the account, they can access the associated Microsoft 365 mailbox, SharePoint folders, or files in the user's OneDrive." For a sales team, that list covers the whole pipeline. Proposals sit in OneDrive. Pricing sheets sit in SharePoint. Every client thread sits in the mailbox.
So the order matters. Cut off access first. Reset the credentials second. Then check what the account reached. Each step depends on the one before it. Ask IT to confirm access is cut off before moving to the reset. If you investigate while they can still send mail, they can use the time to contact more of your clients.
First: cut off the attacker's access
The first job is to stop the account from being used. This is the step most likely to be delayed, by a meeting, by a phone call, or by a debate over whether the account is really compromised. Do not wait for certainty. A message sent from the wrong person is enough reason to act, and the cost of acting on a false alarm is much lower than the cost of waiting on a real one.
Microsoft's guidance gives a clear starting point. It says that if a mailbox shows signs of compromise, you should use its steps to secure and restore the account and regain control. Microsoft's guidance puts disabling the affected user account at the front of the process, as Step 1. In this case, the practical version is simple. Your IT or security team disables the account; ask them to confirm in writing what the disabled state blocks.
Ask your IT team to confirm two things, not just one. First, that the account is disabled. Second, that any sessions already signed in are ended. Ask the person running the response to confirm in writing that existing signed-in sessions have ended. The technical steps belong to your IT or security team. Your job is to make sure the question was asked and answered.
While this happens, handle the rep and the clients separately. Tell the rep, by phone or in person if possible, not to sign in, not to reply to any suspicious message, and not to delete anything in the mailbox. Deleting messages feels like cleaning up, but it removes evidence your IT team needs. The rep should not use the same device to sign in to other accounts until IT says it is safe.
It also helps to name one person as the owner of the response. In a sales organization, several people will want updates within minutes: the deal owner, the sales director, finance, and sometimes the rep's manager. If each of them calls IT separately, the people doing the technical work lose time answering questions instead of fixing the problem. Pick one channel and one point of contact on your side, and send everyone else there.
Do not send a notice to clients from the rep's mailbox. That is the account the attacker controls. A message from it saying "please ignore the last email" tells the attacker you have noticed and tells the client nothing they can trust. Instead, reach the people involved by a route you already know works, such as a phone number already on file or a message from a colleague's account that IT has checked.
In this case, the procurement contact should not be answered from the rep's mailbox at all. Call the contact using a number from your records, confirm that no bank change was requested, and tell them the invoice details should not be updated based on email alone. This is the one message where a reply could cause real money to move, so it needs the most care.
Write down what you did and when you did it. Note the time you first learned of the problem, the time you asked IT to disable the account, the time it was confirmed disabled, and who you spoke to. You will need this record later, both for the investigation and for any conversation with your legal or compliance contact. Keep it short and factual.
Second: reset the password, and do it in the right place
Once the account is disabled, reset the credentials. The guidance on this is short. It points to the page on resetting passwords when it says: "For instructions, see Reset passwords in Microsoft 365 for business." Your IT team will know which steps apply to your setup.
The place where the password lives is the detail most likely to go wrong. Many companies sign staff in through a directory they run themselves, with Microsoft 365 trusting that directory for sign-in. In that setup, a password change made only in the cloud does not fully fix the problem. Microsoft's guidance covers this case directly: "If the user's identity is federated with Microsoft 365, you must change the account password in the on-premises environment and then notify the administrator of the compromise."
In this case, the first question to IT is whether the rep's identity is federated. If it is, the password change happens in the directory, and then the administrator is told about the compromise. If IT changes the password in the wrong place, the attacker's stolen password may still work, and the team may believe the problem is solved when it is not. Ask IT to tell you, in plain words, where the password was changed and who was notified.
The order of the reset and the re-enable also matters. Microsoft's guidance covers the case where the account was disabled during the investigation: "If you disabled the account during the investigation, reset the password and then enable the account as described in Step 1: Disable the affected user account." Do not let a deadline or a client request push the team to re-enable the account before the reset is done.
The rep may also have used the same password on other services. If so, those need to change too. Ask the rep to list the tools they use for client work, such as a CRM, a calendar, a file-sharing app, and any sales tool that signs in with an email address. Each one that used the same password needs its own reset. This is tedious, and it is easy to miss under pressure, which is why it helps to have the list written down before anyone starts. Keep the new password out of email, chat, and shared documents. If it has to be passed on, pass it through the channel IT chooses, and only to the rep.
Before the account is turned back on, the team should decide how the rep will sign in. The goal is to stop the next attacker from using a stolen password alone. Enable two-step verification on the account before the rep goes back to work. The next section explains where to find setup instructions.
Third: find out what the account reached
With access cut off and credentials reset, the third step is to find out what the attacker did and what they could see. This is the part of the response that takes the longest, and it is where revenue leaders can add the most value by asking good questions.
Microsoft's guidance frames the scope clearly: "Remediation and investigation of a compromised user focuses on the affected account and the services associated with the account." That means the review covers the mailbox, but also any other service tied to the rep's account. In this case, the list of questions looks like this.
First, what did the attacker send? Ask IT for the full sent history for the period when the account was likely in use, and match it against your CRM. Every outgoing message should be linked to a real deal or a real contact. Any message that is not should be treated as a possible phishing attempt, and the recipient should be contacted separately.
Second, did the attacker change how mail behaves? Ask IT to check for inbox rules, forwarding settings, and delegated access that the rep did not create. Ask IT to record what they find and confirm which settings need to be removed. Include that result in the response record rather than assuming the password reset addressed every setting.
Third, ask for the sign-in history as well. Ask IT to state in writing which sign-in details the log records and what it shows about the attacker's access. Ask whether the recorded locations match places the rep has worked from. Your IT team can explain what the log shows.
Fourth, what files could they open? Ask IT to review OneDrive and SharePoint for recent sharing activity, especially links created by the account or files moved or downloaded in the same window. In this case, the pricing sheet for the open deal sits in a shared folder. If IT can show the sheet was never opened from an unusual location, that is good news. If the logs are unclear, treat the sheet as exposed and tell the deal team, so they can decide whether the pricing needs to change.
Fifth, who needs to hear about it? The clients who received the suspicious messages need to know, and the procurement contact needs to know not to act on the invoice request. Whether anything more formal is required depends on what was sent, what data was in the mailbox, and where your company operates. Those are questions for your legal or compliance contact, not for the sales team to decide alone. This post cannot tell you whether a notice is required in your case, so ask the question early, while the investigation is still running.
Sixth, what other accounts need two-step verification today? Check every rep's account, and any shared sales account. The NCSC's guidance for small organisations on securing important online accounts says: "Below we've provided links that give specific instructions about turning on 2SV for common products." The page links to setup instructions for common products, which is a practical place to start if your IT team wants a reference. Turn it on for every account that touches client work, not only the email account.
Finally, close the loop with the rep and the rest of the team. A rep whose mailbox was used against clients may feel responsible, even when the fault lies with the attacker. Say plainly that a compromised account is not a personal failure. Ask the rep to report anything unusual they notice, and tell them who to contact. People who feel blamed hide problems. People who feel supported report them early, and early reports are what stop the next compromise from reaching a client.
Keep leadership informed
The response also needs a short, honest update for the leadership team while the work is underway. Keep it to what is known, what has been done, and what is still open. Avoid guessing at the cause or the number of affected messages until IT has the logs. A status note that says "the account is disabled, the password reset is in progress, and the review of sent mail is open" is far more useful than a confident story that changes the next day.
The three steps look obvious when they are written down. In a real morning, they are easy to tangle. A client is asking for an answer, a senior leader wants to know the damage, and someone wants to start the forensic review before the account is even locked. The order exists to prevent that tangle.
Illustrative example: track the review
The team checks 2 against the CRM, leaving 8 to review. Record those remaining messages as open questions, rather than reporting that the investigation has finished. For the invoice-change message, call the procurement contact using a number from your records and record the conversation separately. These counts are illustrative, not a deadline or a measure of the attack's full reach.
Take the next step today
Name your response owner today and ask IT where the rep's password must be reset. Put the owner's contact route and IT's answer in your response record so the team knows whom to call.
Common questions
Where should a federated account's password be changed?
Change it in the on-premises environment and notify the administrator of the compromise.
Who decides whether formal notice is required?
Your legal or compliance contact should assess whether formal notice is required for your case.