MFA Is Not One Control: Choosing Phishing-Resistant Sign-In

Many businesses can truthfully say they have enabled multi-factor authentication and still remain vulnerable to account takeover.

The reason is simple: MFA is a category, not a single level of protection.

A texted code, a push notification, an authenticator-app code, and a FIDO security key all add protection beyond a password. They do not resist the same attacks.

For business leaders, the decision should not be “MFA or no MFA.” It should be:

Which authentication method is appropriate for each user, application, and business consequence?

Why ordinary MFA can still fail

Some authentication methods ask users to manually enter a code. An attacker operating a convincing fake login page can capture that password and code, then relay both to the legitimate service.

Push notifications can also be abused through repeated prompts or social engineering. A user may approve a request simply to stop the interruption.

NIST’s current definition of phishing resistance focuses on preventing authentication secrets or valid outputs from being disclosed to an impostor verifier without depending on the user noticing the deception. It explicitly says manually entered one-time codes are not phishing-resistant because they are not bound to the specific authentication session. pages.nist.gov

This does not mean conventional MFA is worthless. CISA says any MFA is better than none, but describes phishing-resistant MFA as the preferred standard. cisa.gov

What phishing-resistant authentication changes

FIDO/WebAuthn authenticators—including appropriately implemented security keys and passkeys—use public-key cryptography and bind authentication to the legitimate website.

Instead of giving the user a reusable secret to type, the authenticator proves possession of a private key associated with that service. A fake website cannot simply collect and replay a valid code.

NIST finalized Digital Identity Guidelines Revision 4 in July 2025. The update integrates syncable authenticators such as passkeys, and its AAL2 requirements say services must offer at least one phishing-resistant option. pages.nist.gov

That is a meaningful signal for private businesses even when the federal requirements do not apply directly to them.

Prioritize by business impact

A company does not necessarily need to migrate every account on the same day. Begin where an account takeover would have the greatest consequence:

  1. Identity-platform and directory administrators

  2. Cloud, security, and infrastructure administrators

  3. Email and collaboration administrators

  4. Finance, payroll, and payment approvers

  5. Executives and executive-support staff

  6. Software-development and source-code administrators

  7. Customer-support users who can reset identities

  8. Third parties with privileged or persistent access

This risk-based sequence reduces the most dangerous exposure first.

Passkeys and security keys involve tradeoffs

Phishing-resistant authentication is strong, but implementation still requires business decisions.

Synced passkeys can improve usability across a person’s devices. Their security also depends on the protections and recovery process of the synchronization account.

Hardware security keys can offer strong isolation and are useful for highly privileged users. They introduce purchasing, inventory, replacement, and remote-worker logistics.

Device-bound credentials can support managed-device policies but need careful handling during device replacement and disaster recovery.

NIST recommends controls around sync-fabric access, recovery, notifications, multiple bound authenticators, central identity management, and periodic review of registered keys. NIST Pages

The right answer may therefore differ between a customer portal, a managed employee laptop, and a cloud administrator.

Treat recovery as part of authentication

A secure sign-in method paired with a weak help-desk reset process is not a secure identity system.

Before rollout, define:

  • How users enroll their first authenticator

  • Whether a second authenticator is required

  • How lost devices are reported

  • What evidence the help desk requires

  • Which recovery actions need secondary approval

  • How recovery events are logged and alerted

  • How former employees and vendors are deprovisioned

  • How registered authenticators are reviewed and revoked

Test these procedures with realistic scenarios. Attackers frequently target the exception path because organizations focus only on normal login.

A practical rollout plan

Phase 1: Inventory

Identify the identity providers, business applications, privileged roles, existing MFA methods, unmanaged exceptions, service accounts, and recovery processes.

Phase 2: Define tiers

Create a simple authentication standard:

  • Phishing-resistant authentication required for privileged and high-impact access

  • Strong MFA required for the broader workforce during transition

  • Documented, time-limited exceptions with named owners

  • Separate controls for non-human identities

Phase 3: Pilot

Use a representative group spanning executives, administrators, remote employees, support personnel, and multiple device types. Measure enrollment failures, recovery requests, application compatibility, and user friction.

Phase 4: Deploy

Roll out by risk rather than alphabetically. Provide clear instructions, spare authenticators where appropriate, and a verified support process.

Phase 5: Monitor

Track enrollment coverage, fallback usage, recovery activity, legacy authentication, repeated failures, dormant accounts, and privileged accounts without the required method.

The business outcome

Phishing-resistant authentication does more than improve a technical control. It can reduce dependence on users identifying perfect imitations of login pages, strengthen privileged access, and create a clearer identity standard for employees and vendors.

The goal is not to buy security keys indiscriminately or announce a passwordless initiative before applications are ready. It is to build an authentication architecture proportional to business risk—with enrollment, recovery, monitoring, and deprovisioning designed as one system.

Up In Mars can help organizations inventory authentication methods, prioritize high-risk access, evaluate passkey and security-key options, and develop a phased identity-security roadmap.

Next
Next

You Cannot Optimize Multicloud Spending Until You Can Explain It