GUIDE · INCIDENT RESPONSE

Compromised account: response playbook after a successful phishing attack

By the end of this guide you will know how to respond when someone else is using an account after a phishing attack. You will understand why a new password alone is not enough, which checks to run on sessions, rules and logs, and how to decide that the case is genuinely closed.

Updated october 2026
7 min read
Altovar Technical Editorial Team

1. Why a compromised account is an incident

A compromised account is an identity being used by someone else. Whoever holds the password, or a token that keeps the session alive, can read email, send messages as if they were the real person, access shared documents and create rules that hide their tracks. The damage is often not immediate, because attackers tend to observe before acting. For that reason, a quiet period should not be read as a sign that there is no problem.

Phishing is the most common way of obtaining credentials: a message or page that imitates a well-known service and asks you to sign in. Even a careful person can fall for it, because the message seems to come from a supplier or a colleague. The response must therefore start from the assumption that the account has been used, until proven otherwise.

2. What you need before acting

The playbook only works if the people responding have the right authority without needing to ask for permission in the middle of an emergency. You need an administrator of the email and identity platform, with rights to reset credentials, revoke sessions and review access logs. You also need a way to contact the affected user that does not go through the compromised account, for example a phone number or another address already known to you. Whoever has the attacker can read that user's email, so choose the channel with care.

It helps to know which systems are linked to the identity: email, shared files, the management system, signature tools. An account that gives access to many applications needs a broader response than one used only for email. A simple, up-to-date map, with the owner of each system, avoids discovering dependencies in the middle of an incident. The map should be updated whenever new applications or new collaborators are added.

  • An administrator with authority over identity and email
  • A list of the applications linked to the identity
  • An alternative contact channel for the user
  • A written password reset procedure
  • The IT supplier's out-of-hours contact

3. The steps, in order

The playbook starts from the moment someone reports a possible unauthorised access. Each step has a specific goal: block the access of whoever has compromised the account, understand what they did, and give control back to the person. The order matters, because resetting the password without revoking sessions leaves the door open. Communication with the user also needs to follow the right sequence, without asking them to reply to messages sent from the compromised account.

Whoever carries out the steps must record every action with time and name, so the sequence can be reconstructed when it needs to be shown to a consultant or an authority. Notes are worth more than the memory of someone who worked under pressure. Write them down as you act, not at the end of the day, because details fade quickly and some timings cannot be reconstructed later.

  1. Report the anomaly to the technical manager and open the action log.
  2. Temporarily block the account's sign-in, if the platform allows it, without deleting the account.
  3. Revoke all active sessions and linked tokens, including those used by third-party applications.
  4. Reset the password and require a new multi-factor authentication method, that is, a second check in addition to the password.
  5. Check forwarding rules, mail filters and delegated access, because attackers use them to stay hidden.
  6. Review the access logs and the emails sent from the account in the preceding weeks.
  7. Warn contacts who may have received suspicious messages and ask them not to open attachments or links from the account.
  8. Assess with the consultant whether the access involved personal data or confidential documents.

4. Mistakes that let the attacker back in

The first mistake is to reset the password and consider the case closed. If the attacker has already created a forwarding rule or connected an application, they will keep reading email even with the new password. For this reason, revoking sessions and checking delegations should always be done, not only when it seems necessary. The password is only one of the keys to access, and it is often not the most important one.

The second mistake is to delete suspicious emails without recording them. They are needed to understand which message led to the compromise and which other users have been targeted. Improvised communications with the attacker should also be avoided. The third mistake is to assume that an account used for a short time has done no harm: the observation period is often exactly when conversations are studied.

5. How to check that the account is secure again

Verification is not limited to checking that the password works. You need to confirm that no sessions, rules or connected applications left behind by whoever acted remain. The technical manager must be able to state, in writing, which checks were carried out and with what result. A check done only verbally is not enough, because after some time nobody remembers precisely what was verified.

A few days later, repeat the checks, because some delegations may reappear if the attacker obtained broader privileges. Then go back to the person involved for a final review, asking whether they have noticed anything strange or unexpected messages. If they report something unusual, the cycle starts again from the first step, without treating the case as closed. Record the closure with date and owner, together with the evidence collected.

CheckWhat to look forExpected result
Active sessionsSign-ins from unrecognised devices or locationsOnly the user's own sessions
Forwarding rulesCopies of messages sent to external addressesNo unauthorised rules
Connected applicationsAccess granted to unknown softwareOnly approved applications
Sent messagesEmails the user did not writeNo unusual sending

6. How to get started

This week, check whether your email administrator can revoke a user's sessions without going through another supplier. If the answer is unclear, it is the first thing to settle. Also ask how quickly a password reset can be carried out and who can authorise it out of hours. These details may seem secondary, but they become decisive at night.

Then write the steps above on a page, with the name of the person who carries out each one and the number to call out of hours. Run a test with the technical manager, even just on paper, to find where access or information is missing. Keep the document somewhere accessible even if the company email is compromised, for example as a printed copy or in a separate archive.

Frequently Asked Questions

If an employee entered their password on a fake page, is it enough to change it?

No. The new password is necessary but not sufficient. The attacker may already have an active session or have created forwarding rules and connected applications that remain operational. Revoke sessions, check delegations and review the access logs before treating the case as closed.

How do I know whether my account has really been used by others?

The clearest signs are sign-ins from unrecognised locations or devices, sent emails the user does not remember writing, forwarding rules that were never created, and unusual verification requests. The logs of the email and identity platform are the main source. If you cannot read them, ask your IT supplier to extract and preserve them straight away.

Do I need to warn customers if the account contained their data?

It depends on what data was accessible and to whom. The GDPR, the European data protection regulation, may require notification to the supervisory authority within 72 hours of becoming aware of the breach, where a risk to individuals is likely, and communication to the individuals concerned if the risk is high. Assess this with a consultant before writing to anyone.

Is it enough to turn on multi-factor authentication?

It is one of the most effective measures for reducing risk, but on its own it does not close an incident that has already happened. If the attacker already has an active session or rule, the second factor will not stop them. Turn it on for all accounts as prevention, alongside the response procedure and not in place of it.

Can my email provider help me investigate?

It can provide logs, revocation tools and sometimes technical support, but the arrangements vary by service and contract. Do not assume the provider will carry out a full forensic analysis. First check which logs it keeps, for how long and how they can be exported, because retention may be limited.

About this content

General informational guide. Regulations and technical requirements change over time: always check the official sources and assess your specific case with a qualified technical or legal advisor.

GET IN TOUCH

Talk to our team.

We'll show how Altovar integrates into your infrastructure. No commitment, no sales pitch. Just a technical conversation.

BOOK A DEMO →