MFA was designed so that two things have to be true before access is granted. In practice many of the attacks we handle get past one of those things simply by waiting until a tired user taps “approve” late at night — because they were told to trust a legitimate-looking prompt.
Where the fatigue actually comes from
- Too many prompts for low-risk, routine actions
- Prompts with no context — a user cannot tell a real change from an attack by looking at them
- No visible record of what was approved, so anomalies never get flagged
The fix is less about replacing MFA and more about trimming the number of prompts that trigger it. Phishing-resistant flows that only fire for genuinely sensitive transitions — a login from a new location, approval of a transfer, a change to an admin role — keep the strong signal high.