Insights · Microsoft 365 Security

Conditional Access vs MFA: what's the difference?

We get asked this most weeks, usually as "we've got MFA on, so we're covered, aren't we?" MFA is essential. It is also no longer enough on its own, and the gap between the two is where most of the Microsoft 365 account takeovers we get called in to clean up actually happen.

Here is the difference in plain terms, what each one stops, and how we deploy Conditional Access without locking a business out of its own tenant.

MFA is a check. Conditional Access is the rulebook.

Multi-factor authentication asks for a second proof that you are you: a code, an app prompt, a fingerprint, a security key. It answers one question, "is this really this person?"

Conditional Access is the policy engine in Microsoft Entra ID that sits in front of every sign-in and decides what should happen. It looks at who is signing in, which app they want, what device they are on, where they are, and how risky the sign-in looks. Then it makes a decision: allow, ask for MFA, insist on a managed device, or block outright.

So MFA is one of the things Conditional Access can demand. Conditional Access is how you decide when to demand it, and what else has to be true.

Why MFA on its own stopped being enough

Attackers adapted. Three techniques get around a tenant that has MFA switched on and nothing else:

  • Token theft. A convincing fake sign-in page sits between the user and Microsoft. The user types their password, approves the MFA prompt, and the attacker walks away with the session token. MFA happened. The attacker is in anyway.
  • Prompt fatigue. Repeated approval requests sent until someone taps Approve to make them stop. Number matching in Microsoft Authenticator has made this harder, but it has not gone away.
  • Legacy sign-in routes. Old email protocols cannot do MFA at all. If they are still open, a stolen password is all an attacker needs.

The uncomfortable part: in a token theft attack the user did everything they were trained to do. They approved a prompt they were expecting. More training does not fix that. What reduces the risk is a policy that checks where a sign-in is coming from, and refuses it when it is not one of your devices.

What Conditional Access adds

  • Require a compliant device. Sign-ins to company data only succeed from devices you manage and that meet your security rules. This makes a stolen password or session far harder to use from an attacker's own laptop. It reduces the risk of token theft; it does not remove it. For the strongest protection, pair it with phishing-resistant sign-in such as passkeys, and Microsoft's token protection where your licensing supports it.
  • Block legacy authentication. Closes the routes that cannot do MFA.
  • Stronger rules for administrators. Phishing-resistant sign-in methods, such as security keys or passkeys, for the accounts that can do the most damage.
  • Location and risk rules. Block sign-ins from countries you never work in, and step up or block sign-ins Microsoft flags as risky.
  • Session controls. How long a sign-in lasts, and what an unmanaged device is allowed to download.

Security defaults, per-user MFA and Conditional Access

Microsoft gives you three ways to turn MFA on, and they are not equal:

  • Per-user MFA is the old method: a switch per account. No conditions, easy to miss people, and Microsoft is steering everyone away from it.
  • Security defaults are free and on by default in new tenants. Everyone registers for MFA and legacy authentication is blocked. It is a sensible minimum, but it is all or nothing: no exceptions, no device rules, no tuning.
  • Conditional Access needs Microsoft Entra ID P1, which is included in Microsoft 365 Business Premium. Risk-based policies need P2. This is one of the main reasons we move clients from Business Standard to Business Premium: the licence that adds Conditional Access also adds Intune, which is what makes the "compliant device" rule possible.

How we roll it out without locking anyone out

Conditional Access is powerful enough to lock every user, including the administrators, out of the tenant. The order matters:

  • Break-glass accounts first. Two emergency admin accounts, excluded from the policies, with long random credentials stored offline and alerts on any use.
  • Report-only mode. Every policy runs silently first, so we can see who it would have blocked before it blocks anyone.
  • A pilot group. A handful of real users for a week or two, then the rest of the business in stages.
  • Devices before device rules. Laptops and phones are enrolled and compliant in Intune before any policy starts requiring it.
  • Exclusions that are written down. Service accounts and awkward legacy systems are handled deliberately, reviewed, and not left as a growing list of exceptions.

The mistakes we find when we audit a tenant

  • Per-user MFA and Conditional Access both running. Confusing prompts, and nobody is sure which one is actually doing the work.
  • The office marked as a trusted location that skips MFA. Anyone on the guest Wi-Fi, or on a compromised PC inside the building, gets a free pass.
  • One policy for everyone, including admins. Global administrators signing in with the same text-message code as everyone else.
  • No break-glass account. Fine until the day the MFA service has a bad morning, or the only admin loses their phone.
  • Policies switched on and never looked at again. New apps, new starters and new devices drift outside the rules.

So which do you need?

Both. MFA on every account is the baseline, and it is a requirement for cloud services under Cyber Essentials. Conditional Access is what turns it from a single check into a set of rules that still holds when someone approves the wrong prompt.

Not sure what your tenant is actually enforcing?

Read about our Conditional Access consultancy and Microsoft 365 security work, see what Intune does for the device side, or request a free IT review and we will tell you where you stand.

Independently reviewed by our customers