A phone buzzes once during dinner. Then again. And again.

If you did not start a sign-in, do not approve the request. Repeated authentication prompts can be a deliberate tactic called MFA fatigue, or push bombing. The attacker may already have a valid password and is trying to get past the second step by wearing down the person who receives the prompts.

Multi-factor authentication (MFA) remains an important account-security control. The goal is to make sure the second factor is difficult to trick or misuse—and that people know exactly what to do when a request is unexpected.

What is an MFA fatigue attack?

With push-based MFA, a service sends an approval request to an enrolled device when someone tries to sign in. In an MFA fatigue attack, an attacker triggers repeated requests and hopes the user will approve one by mistake, to stop the interruptions, or because a message or caller pressures them to do so.

CISA describes this technique as “MFA fatigue” or “push bombing.” Receiving a prompt alone does not prove an account has been compromised, but it is a warning worth reporting. The request may mean someone is trying to sign in with the account’s credentials.

How push bombing works

A typical sequence looks like this:

  1. An attacker obtains or guesses a user's password, for example through phishing or reuse of a leaked credential.
  2. They attempt to sign in to a service that sends push approval requests.
  3. The legitimate user receives a prompt they did not initiate.
  4. The attacker sends more requests or contacts the user with a pretext, hoping for an approval.
  5. If the request is approved, the attacker may gain access, depending on the service and any other security checks.

This is why an MFA prompt is not a routine notification to tap without thought. It is an authorization decision.

What to do when an unexpected prompt appears

If you receive a login request you did not start:

  • Deny the request. Never approve a prompt just to make repeated notifications stop.
  • Do not share a one-time code with anyone who calls or messages you, even if they claim to be IT support.
  • Report it promptly through your organization’s established security channel. Include the time, service, and whether prompts continued.
  • If you approved one by mistake, contact IT or security immediately. They can revoke sessions, reset credentials, review sign-in activity, and decide whether other response steps are needed.
  • Use a known route to contact support. Do not rely on contact details supplied in an unexpected call or message.

For personal accounts, use the provider’s official app or website to review recent activity, change a potentially exposed password, and sign out other sessions where the service offers that option.

How organizations can reduce the risk

Enable number matching as an interim safeguard

Number matching requires the person signing in to enter or select a number shown in the sign-in flow. This extra step makes blind approval of a stream of prompts harder. CISA recommends number matching for organizations that cannot yet implement phishing-resistant MFA.

Number matching helps reduce push-bombing risk, but it does not make every sign-in phishing-resistant. Follow the identity provider’s current configuration guidance and test the user experience before broad rollout.

Move toward phishing-resistant MFA

CISA identifies FIDO/WebAuthn and public-key-based methods among phishing-resistant options. NIST defines phishing resistance by whether the authentication protocol prevents authentication secrets or valid outputs from being disclosed to an impostor verifier, without relying on the user to spot the deception.

For many teams, this means evaluating passkeys or security keys that use FIDO/WebAuthn, where the identity provider and applications support them. Plan enrollment, account recovery, lost-device handling, and support for people who cannot use a particular device. Keep recovery paths protected too: a strong sign-in method can be undermined by a weak reset process.

Make prompts understandable

Tell users what a legitimate approval request looks like and make it easy to report an unexpected one. Explain that support staff should not ask them to approve a sign-in they did not initiate or to read out an authentication code.

Prepare an account-response process

Security and IT teams should know how to quickly review sign-in logs, revoke sessions, reset credentials, and confirm the user has regained control. The exact steps depend on the identity provider and the organization’s incident-response procedures. Practice the workflow before an incident.

Review authentication coverage

Inventory critical email, administrator, remote-access, and business applications. Confirm which MFA methods are enabled, which accounts are exempt, and how recovery works. Prioritize high-impact accounts and remove exceptions that no longer have a clear owner or reason.

Push approval, number matching, and phishing-resistant MFA

  • Push approval: The user taps approve or deny. This is convenient, but repeated prompts can pressure users into approving one.
  • Push with number matching: The user enters a number displayed in the sign-in flow. This helps mitigate push bombing, but it is not phishing-resistant.
  • FIDO/WebAuthn or passkey: The user authenticates with a device-bound or synced credential through a supported authenticator. This is designed to resist phishing by binding authentication to the legitimate service.

Exact behavior varies by identity provider and configuration. Check the provider’s documentation before changing authentication settings.

A short response checklist

If someone reports repeated MFA requests:

  1. Confirm they did not initiate the sign-in and ask whether they approved any request or shared a code.
  2. Follow your identity provider’s incident process to review activity and secure the account.
  3. Revoke sessions and reset credentials when indicated by your investigation.
  4. Check whether other accounts or applications show related activity.
  5. Document the event and close gaps in MFA configuration, recovery, or user guidance.

Frequently asked questions

Is MFA fatigue the same as phishing?

No. MFA fatigue specifically uses repeated authentication prompts to pressure a person into approving a sign-in. It can happen alongside credential theft or social engineering, but the terms describe different parts of an attack.

Should I approve a prompt if an IT employee asks me to?

Only approve a sign-in you initiated yourself. If someone says they need you to approve a request, stop and verify through your organization’s known support channel.

Is SMS MFA a good replacement for push notifications?

Changing MFA methods requires a review of the risks, available options, and provider support. CISA ranks phishing-resistant methods strongest and notes that SMS or voice methods have additional weaknesses. Follow current guidance for your environment rather than switching methods without a plan.

Does phishing-resistant MFA stop every account takeover?

No single control prevents every attack. Phishing-resistant MFA addresses credential theft through impostor sign-in sites, but organizations still need protected recovery processes, monitoring, and incident response.

Key takeaways

  • An unexpected MFA request should be denied and reported.
  • Repeated prompts can be a sign that someone is trying to use the account.
  • Number matching can reduce accidental approval, but it is an interim safeguard.
  • Organizations should plan for phishing-resistant MFA such as FIDO/WebAuthn where supported.
  • Protect account recovery and make it easy for users to report suspicious requests.

For authoritative guidance, see CISA’s number-matching fact sheet, CISA’s phishing-resistant MFA fact sheet, and NIST SP 800-63B-4.