MFA Won’t Stop Credential Stuffing, But Zero Trust Can

Cyberpunk-style Doberman in tactical gear next to a suitcase full of fake FBI credentials and passports, illustrating that MFA won’t stop credential stuffing.

MFA Is Not a Catch-All

You’ve implemented MFA. Maybe you’ve even rolled it out org-wide and added some baseline conditional access. On paper, your identity security posture looks strong. But if credential stuffing hasn’t been addressed at the foundation, you’re still exposed.

Credential stuffing attacks rely on one simple fact: users reuse passwords, and those passwords are already out there, breached, sold, and packaged up for automation. Attackers don’t need to guess; they just need to log in. And if legacy protocols are still enabled, or session tokens aren’t protected, MFA won’t stop them.

At Levacloud, we work with teams like yours who’ve already invested in Microsoft’s security stack, but haven’t yet hardened the parts attackers are quietly targeting. This blog will show you why credential stuffing continues to work, what MFA isn’t catching, and how Zero Trust gives you the control you actually need.

What Is Credential Stuffing, Really?

Credential stuffing is exploitation of valid username and password pairs gathered from past breaches, malware infections, or purchased from dark web marketplaces. The attackers automate login attempts at scale, targeting anything from VPN portals and webmail to legacy apps and SaaS logins.

This technique works because people reuse passwords, and because many systems still accept logins with little friction. Even if MFA is technically “enabled,” that doesn’t mean it’s enforced everywhere or configured to catch this behavior.

If your organization is using Microsoft Entra ID but hasn’t implemented adaptive access policies, token protection, or full coverage of legacy protocol blocks, you’re still vulnerable. We see this all the time when we run assessments at Levacloud. Credential stuffing isn’t flashy, but it’s still one of the top ways attackers get their foot in the door.

Sign Up To Our Newsletter

We’ll keep you up to date on the latest in Microsoft Cybersecurity.

Why Credential Stuffing Still Works in MFA-Enforced Environments

On the surface, MFA looks like it should shut down credential stuffing completely. But in practice, attackers bypass it all the time because of how it’s deployed, not because the technology itself is flawed. Here’s how it happens:

Inconsistent MFA Coverage

Credential stuffing thrives when MFA isn’t enforced on every app, user, or authentication method. We still see enterprises where service accounts, legacy apps, or low-priority users are excluded from policies, giving attackers exactly the entry point they need.

Weak MFA Methods

MFA that relies on SMS or email is easy to intercept or socially engineer around. Even push notifications without number matching are vulnerable to fatigue attacks, where attackers bombard the user until they tap “approve” just to stop the noise.

Session Token Replay

Once attackers log in successfully, even just once, they can often extract and reuse session tokens to bypass MFA on subsequent access. Without token protection in Entra Conditional Access, there’s nothing stopping session hijacking from working.

Legacy Protocols Still Enabled

IMAP, POP3, SMTP AUTH, these older protocols don’t support modern MFA. If they’re still turned on (even for just one system), they give attackers a clean path in using only a username and password.

Help Desk and Social Engineering Gaps

We’ve worked with enterprise clients who had strong MFA policies in place, on paper. But their help desk processes allowed attackers to socially engineer their way into registering a new MFA device on an executive account.

At Levacloud, we don’t just check the MFA box, we dig into how and where it’s enforced, where it be exploited, and how to close loopholes with policy, not just software. You already have the right tools. We help you configure them so they actually protect you.

Not sure if your MFA is Credential Stuffing Proof?

Discover how Levacloud can help close your security gaps.

How Credential Stuffing Leads to Bigger Breaches

Once an attacker gains access to a valid account, especially in an enterprise environment, they aren’t just interested in email or documents. They’re looking for lateral movement opportunities, sensitive data, elevated permissions, and paths to deploy malware or ransomware. And if the compromised account belongs to a user with wide access, or if internal segmentation is weak, that attacker can pivot quickly.

Some of the most costly breaches start with nothing more than a reused credential from a public breach.

Credential stuffing is notoriously hard to detect once an attacker is in. There’s no malware, no exploits, just a valid login that doesn’t trigger many alarms unless your identity tools are configured to look deeper. That’s why session analytics, sign-in risk scoring, and behavioral anomaly detection are critical, but often overlooked.

Zero Trust For Credential Stuffing Defense

You probably know that Zero Trust is not a product. It’s a security model built on a simple principle: never trust, always verify. That means identity, device health, location, session behavior, and risk signals are all evaluated continuously, not just at the moment of sign-in.

When properly implemented, a Zero Trust architecture stops credential stuffing attacks from moving beyond that first login attempt with multiple layers.

Most organizations already have elements of Zero Trust in place through Microsoft Entra ID, Defender for Identity, and Intune. But having the tools isn’t the same as having them configured to stop attacks like credential stuffing.

Are You Dealing With A Microsoft Cybersecurity Challenge?

You have a pressing issue, but you’re not sure if Levacloud can help. We get it. Everyone has unique challenges they face in their IT environments. Schedule a free call today and talk us through it.

We’ll let you know how we can best support you.

Ways To Harden Your Security Against Credential Stuffing

1. Block Legacy Authentication Completely

Credential stuffing succeeds when legacy authentication protocols are still enabled. These include IMAP, POP3, SMTP AUTH, and MAPI over HTTP, which allow basic authentication and ignore Conditional Access and MFA enforcement.

Even if you’ve rolled out modern MFA across the board, any user or system authenticating through a legacy protocol is still exposed. Attackers take leaked credentials from past breaches and spray them across public-facing hoping one hits a legacy path where MFA doesn’t apply. Here’s what to do:

  • Audit usage first. In Entra ID, go to Sign-in logs → Client App filter to identify which users and service accounts are still authenticating using basic protocols. You’ll likely see a long tail of outdated mail clients, legacy line-of-business apps, or misconfigured third-party tools.
  • Set a deadline. Microsoft will fully disable basic authentication in Exchange Online by September 2025, but don’t wait. Begin migrating off now, and treat any remaining usage as a critical risk.
  • Block at the protocol level. Use Conditional Access rules targeting client app types to enforce protocol-specific blocks. This gives you granular control while avoiding false positives.
  • Segment and isolate service accounts. If legacy auth is still required for a legitimate use case (e.g., copier scan-to-email), restrict it by static IP range, isolate it in its own group, and disable interactive sign-in.

2. Enforce Strong MFA Everywhere

Credential stuffing becomes harder to pull off when strong, phishing-resistant MFA is enforced across all accounts and applications. Don’t fall into the trap of thinking that simply enabling MFA in Entra ID is enough. It’s not. The common pitfalls we see:

  • MFA is enabled, but not required due to lax Conditional Access scope.
  • It’s required only for “admins” or specific groups.
  • Users authenticate via methods that are easy to exploit, like SMS codes or app-based OTPs that can be intercepted or phished.
  • Service accounts, break-glass accounts, and older SaaS integrations are often excluded entirely.

To close these gaps, your approach needs to be deliberate, not default.

  • Use Conditional Access to enforce MFA at the point of sign-in, don’t rely on registration or app policies alone. Make sure all user types (contractors, guest users, execs, service owners) are in scope.
  • Move away from weak methods. Disable SMS and voice call MFA where possible. Favor stronger methods:
    • Microsoft Authenticator with number matching
    • FIDO2 hardware keys
    • Passkeys bound to compliant devices
  • Apply layered controls. For example, use MFA and device compliance for admin accounts. Require multiple signals before granting access to sensitive apps (e.g., user risk, device trust, IP location).
  • Audit for stale or orphaned registrations. Use Entra reporting to find users with registered MFA methods that are no longer in use, or never completed MFA setup.

If you’re not confident that your MFA policies cover every entry point, or if you’re unsure how to handle exceptions like third-party apps or older services, our team can help. We can help you ensure strong methods are enforced without breaking operational workflows.

3. Implement Risk-Based Conditional Access

Credential stuffing works because attackers use valid credentials from real users, often from normal-looking devices, at normal times, using familiar endpoints. That’s why static Conditional Access rules fall short. You need policies that adapt to context.

Risk-based Conditional Access leverages signals from Microsoft’s threat intelligence to block or challenge logins that look legitimate but act abnormally. This is what turns MFA from a gate into an ongoing evaluation. What to configure:

  • Enable sign-in risk and user risk policies. These are built into Entra ID P2 and evaluate signals like:
    • Impossible travel (e.g., logging in from Virginia and Vietnam 10 minutes apart)
    • Use of anonymizing services (e.g., TOR, residential proxies)
    • Known compromised credentials or leaked tokens
    • Behavioral anomalies (time-of-day patterns, first-time geo, etc.)
  • Set real actions for risk levels. For high-risk sign-ins, configure policies to:
    • Block the request outright
    • Require phishing-resistant MFA methods
    • Trigger a password reset
    • Require a compliant or hybrid-joined device
  • Combine risk with device and location context.
    • Allow sign-ins only from known corporate IP ranges
    • Require compliant devices for access to sensitive apps
    • Apply stricter rules to accounts with admin privileges, external access, or access to finance systems

What’s often missed:

  • Many Conditional Access environments are policy-fragmented, built incrementally with overlapping rules and exceptions.
  • Risk signals may be disabled or only used in audit mode.
  • No process in place to respond to user or sign-in risk alerts.

This is a common place where enterprise environments hit a wall: you’ve got all the knobs, but you’re unsure how to turn them without causing disruption. We can help you build Conditional Access frameworks without creating policy sprawl or false positives that lead to alert fatigue.

4. Protect Session Tokens with Token Binding

Even when MFA is enforced and credentials are verified, attackers who gain access to a valid session token can bypass every control you’ve put in place. That includes Conditional Access, device trust, and even phishing-resistant MFA, because the authentication has already happened.

This is why token replay attacks are so dangerous. Credential stuffing followed by session hijacking is becoming more common, especially with token-stealing malware like Redline, Vidar, or browser-injection kits that extract tokens from memory or local storage. Here’s what to do:

  • Enable Token Protection in Entra Conditional Access.
    This feature binds the access token and refresh token to the user and their device. If an attacker attempts to replay the token from another machine, the token is invalidated, regardless of whether it’s technically still “valid.”
  • Set enforcement for privileged roles first.
    Start with highly targeted groups, Global Admins, Entra ID PIM users, Exchange Admins, etc., then expand to broader populations as compatibility testing completes.
  • Use compliant or hybrid-joined devices where possible.
    Token protection works best when combined with Intune-managed or hybrid-joined devices that can provide strong device identity signals.
  • Pair it with continuous sign-in evaluation.
    This revokes tokens if risk conditions change after login, such as if a device moves locations, shows new risk signals, or becomes non-compliant.

What’s often missed:

  • Many enterprises don’t know that token binding is opt-in, even if they have Entra ID P2.
  • Service accounts, unattended scripts, and browser sessions often fall outside enforcement unless scoped properly.
  • Some third-party apps or browser extensions interfere with token binding and need to be tested in advance.

If you’re unsure how token replay might show up in your environment, or if you’re hesitant to enable Token Protection due to compatibility questions, we’ve already walked multiple enterprise teams through it. At Levacloud, we help scope, test, and roll out token protections in a way that’s safe, staged, and monitored closely through audit logging and sign-in analytics.

5. Deploy Microsoft Defender for Identity

Credential stuffing is just the entry point. Once an attacker is in, they start moving laterally and probing for ways to escalate privileges.

Microsoft Defender for Identity (MDI) is designed to catch what identity-based attacks look like after a login has succeeded. It monitors Active Directory domain controllers for behavioral anomalies that signal compromise, especially in hybrid environments where cloud identity is tied to on-prem infrastructure. Here’s what it detects:

  • Lateral movement techniques like Pass-the-Hash, Pass-the-Ticket, and Overpass-the-Hash
  • Reconnaissance behavior like directory enumeration, DCSync attempts, and Kerberoasting
  • Privilege escalation attempts through group membership changes, GPO edits, or Golden Ticket usage
  • Abnormal behavior from valid accounts, logins at unusual times, privilege use from new endpoints, or high-volume LDAP queries

What to do:

  • Deploy MDI sensors to every domain controller (physical or virtual). These lightweight agents capture domain traffic in real time, without needing SPAN ports or mirror traffic.
  • Integrate with Microsoft 365 Defender.
    This unifies your signal across endpoints (Defender for Endpoint), email (Defender for Office 365), and identity (MDI), giving you cross-domain correlation and automatic incident creation.
  • Tune alerts and create investigation workflows.
    Out-of-box alerts are strong, but noisy in large or complex environments. Suppress known service accounts or legitimate tools, and set up automated investigations where appropriate.
  • Pair with Sentinel or SIEM for visibility.
    Forward alerts and raw data to Microsoft Sentinel or your SIEM of choice to enrich your broader detection and response workflows.

If you haven’t deployed MDI yet, or deployed it but haven’t tuned or integrated it, you’re missing key signals that turn credential stuffing into a domain compromise.

6. Educate Staff on Push Fatigue and MFA Exploits

When attackers successfully use stolen credentials in a credential stuffing attack, and MFA is enabled, their next step is often to exploit the human element.

One of the most common methods? MFA push fatigue.

Attackers initiate dozens of login attempts, each triggering a push notification to the user’s mobile device. Sooner or later, especially during off-hours or after multiple attempts, the user taps “Approve” just to make it stop. It only takes one, and it’s not the only tactic:

  • Consent phishing tricks users into authorizing a malicious app with legitimate OAuth scopes
  • Device code phishing uses legitimate Microsoft URLs but prompts users to enter a device code that grants access
  • Help desk social engineering sees attackers impersonating users to reset MFA, change devices, or register new authenticators

Even the best technical controls can’t protect you from an attacker who’s been invited in by an exhausted, tricked, or overwhelmed user. So, what can you do?

  • Enable number matching in Microsoft Authenticator.
    This eliminates one-tap approvals by requiring the user to enter a number shown on the screen, thwarting push fatigue attacks.
  • Block device code sign-ins for users who don’t need them.
    Use Conditional Access to control which user groups are allowed to use device code flow, and always monitor usage.
  • Restrict OAuth consent.
    Use Entra ID admin consent workflows to prevent users from granting high-risk app permissions. Log and review granted OAuth tokens regularly.
  • Train users on MFA threats, not just phishing emails.
    Educate employees on what an unusual MFA prompt looks like, why unexpected approvals should be denied, and who to notify if they’re unsure. Include high-risk roles like execs, HR, and IT admins in targeted awareness efforts.

7. Stop Password Reuse with Banned Password Lists

The unfortunate reality is that most enterprise environments don’t prevent password reuse beyond minimum length and complexity, and attackers know it. If a user’s password was exposed in a LinkedIn, Dropbox, or Adobe breach five years ago, there’s a good chance it’ll still work somewhere today. What can you do?

  • Enable Microsoft Entra Password Protection.
    This allows you to block commonly used, weak, or known-breached passwords, both in Azure AD and on-premises. The Microsoft global banned password list is updated regularly and includes real-world breach data and dictionary-based variants.
  • Extend protection to on-prem AD.
    Deploy the Password Protection proxy and DC agents to your local domain controllers. This lets you enforce banned passwords across hybrid identities.
  • Create a custom banned password list.
    Include organization-specific terms (company name, acronyms, location, season, etc.) to prevent predictable variants. This is especially important for execs and privileged accounts.
  • Monitor password change events.
    Use Entra reporting and SIEM integration to flag repeated failed password resets or bypass attempts, which may indicate a user attempting to use a weak or blocked password.
Ask Us A Question

Wondering if Levacloud can solve your Microsoft Cybersecurity related challenge? Drop us a message!

This field is for validation purposes and should be left unchanged.

Stop Thinking of MFA as the Finish Line

If your organization treats MFA as the end goal for identity security, you’re leaving the door open. Credential stuffing doesn’t rely on exploits, malware, or zero-days, it relies on overlooked defaults, legacy systems, and human habits. And those don’t get fixed with a single setting.

MFA is critical, but credential stuffing still works when:

  • Session tokens aren’t bound to the device
  • Legacy protocols are still enabled
  • Users approve prompts without thinking
  • Passwords are predictable or reused
  • Risk signals aren’t used to adapt enforcement

If you’re unsure whether your environment is hardened against credential stuffing, or you’ve rolled out the tools but haven’t had time to optimize them, we can help.

You’ve done the hard work getting MFA in place. Now it’s time to make sure it actually holds.

Post Reviewed by Gareth Young, CISSP

This blog post was reviewed and validated by Gareth Young, a Microsoft Security and Compliance Expert with 15 years of experience in Microsoft solutions. As the founder of Levacloud, Gareth specializes in Security, Modern Work and Security Arcitecture. He holds multiple Microsoft certifications, including: AZ-500, MS-500, SC-400, MS-101, MS-100, MS-900 as well as the CISSP certification.

Gareth Young
LinkedIn

Related Posts