Cybersecurity

Setting up MFA or user permissions? Separate identity, policy and access.

Adding an authenticator does not grant access to a file or prove MFA is enforced everywhere. Confirm the work account, approved method and actual permission needed before making a change.

16 October 2026  •  9 min read

Woman using a smartphone beside a navy laptop at a pale kitchen counter with green tiles

For authorised account setup and access review

Follow your organisation’s identity and information-handling rules. Individual registration, tenant MFA policy, administrative roles, app consent and file permissions are separate controls with different owners.

Step 1 of 10

In this guide
  1. 01Identity and method setup
  2. 02Policy and permissions
  3. 03Errors, verification and help

Confirm the work account, owner and task

Personal, work and guest identities can have different access.

Confirm the exact work/school account and organisation, the device you are authorised to use and whether you are registering a sign-in method, changing MFA policy or assigning access. A personal Microsoft account with a similar address is not interchangeable with the work identity. A guest may be accessing a different organisation’s resource.

Use the organisation’s known sign-in or support entry point. Check the account and organisation shown before proceeding; do not follow an unexpected email or QR code offering to enrol your account. If prompts or permission changes are unexpected, report them through a known contact rather than treating them as routine setup.

Agree who owns the task. A user can register methods allowed for them, but this does not authorise changing tenant-wide policy, granting administrative roles or sharing somebody else’s records. For an existing sign-in failure, the Microsoft 365 sign-in guide helps distinguish the blocker.

Identify the correct account, organisation, authorised device and task owner. Do not mix personal recovery, work registration, policy enforcement and resource access.

Next: Check current methods, recovery and approved policy

Check current methods, recovery and approved policy

Plan a phone change or rollout before removing working access.

Before adding or replacing a method, confirm which approved methods currently work and how the organisation verifies identity when a phone is lost or unavailable. Keep an existing permitted method available while the replacement is enrolled and verified. Do not delete the old method or reset all security information as a first experiment.

Ask the administrator which authentication methods are enabled for your account and whether a managed device, registration location or other policy requirement applies. A method offered to one account may not be allowed for another. Agree an accessible supported option if the usual phone or method cannot be used; do not enrol someone else’s phone or share one identity.

For an organisational rollout, the responsible administrator needs a staged plan covering affected users, applications, recovery and continuity before changing enforcement. Record the support contact and unresolved dependencies. Do not sign everyone out or switch off existing protection merely to test a proposed setup.

Keep a working approved method while a replacement is verified. Establish the permitted method, support and recovery process before a phone change or policy rollout.

Next: Register the permitted method from genuine Security info

Register the permitted method from genuine Security info

Use the work-account flow and keep enrolment details private.

For an allowed Microsoft Authenticator setup, use the official current app from your phone’s recognised app store. From the organisation’s known account entry point, open the work account’s Security info page. Choose Add sign-in method and the method the organisation allows. If the option is missing or blocked, ask the administrator; do not work around registration policy.

For the official work-account Authenticator flow, open the app, add a Work or school account and scan only the code generated in your own genuine setup session. Follow that session’s verification instructions and confirm the correct account. Never send the enrolment QR code, setup secret, password or one-time code to a caller or include it in a support screenshot. Do not approve a prompt you did not initiate.

The available options and steps can differ for passkeys, security keys and other supported methods. Use the organisation’s specific instructions rather than applying the personal-account flow to a work account. A successfully added account entry is only registration; have the administrator confirm how enforcement applies, then verify the intended task.

Register only an approved method in the genuine work-account flow and finish its verification. Keep enrolment secrets private; an app entry alone does not establish enforced MFA.

More checks7 detailed checks

Separate registration from MFA enforcement

The administrator must check the actual policy and affected scope.

Ask the responsible administrator to inspect the actual tenant configuration: Security defaults, Conditional Access where licensed and configured, and any existing policy or migration dependencies. Registering a phone or installing an app does not itself require MFA for every application and session. A quiet sign-in also does not by itself show MFA is off; an existing session or supported authentication method may already satisfy the requirement.

Security defaults provides preconfigured protections; Conditional Access allows more specific conditions when the required licensing and configuration are in place. Do not toggle either control or create exclusions from a generic guide. The administrator should assess existing protection, legacy clients, affected users and continuity, then use a controlled rollout and review the resulting sign-in evidence.

Authentication method strength matters too. A push approval or code is not automatically phishing-resistant. Microsoft supports phishing-resistant options such as permitted FIDO2/passkey methods, subject to the actual device and policy. Passkeys in Authenticator are different from ordinary Authenticator push approvals. Do not promise that any MFA prevents every compromise or that every sign-in must show a fresh prompt.

Have the policy owner confirm real enforcement, method strength, coverage and continuity. Registration, prompts and a single successful sign-in do not establish tenant-wide protection.

Next: Verify the agreed sign-in and access scope

Identify the resource and the least access needed

File access, administrative roles and app consent are not the same permission.

Write down the actual task and resource: viewing a particular document, editing a team library, accessing a mailbox or administering a service. Confirm the intended person’s exact work or guest identity, required view/edit/administration level, business owner and duration. Being able to sign in does not entitle someone to every resource.

Distinguish a Microsoft 365 licence, group membership, resource permission, Entra/service administrative role and an application’s consent. Each affects different things. Do not assign Global Administrator, approve organisation-wide app access or add someone to a broad group just to open one file. Use least privilege: the permissions needed over the appropriate scope and time.

For a sharing problem, review the actual source and intended audience using the file sync and sharing guide. For a licensed feature, check the licence guide. Neither adding a licence nor registering MFA substitutes for the resource owner’s permission decision.

Specify the exact identity, resource, task, access level, scope and duration. Choose the right control; do not grant broad administration for an ordinary file or app-access request.

Next: Review existing access and use the responsible owner

Review existing access and use the responsible owner

Effective access can come through several routes.

Have the authorised owner review current direct permissions, relevant group membership, inherited access and sharing links for the actual resource. A restricted new link does not remove access already granted another way. Removing a direct permission may also leave access through a group. Record the actual effective access rather than judging by one invitation.

For administrative work, the role owner should choose an appropriate documented role and scope, with time-limited privileged access where the organisation supports it. Protect administrative identities and retain the organisation’s established recovery and emergency-access arrangements. Do not create a new shared admin account, assign a broad role or change emergency-access policy as a shortcut.

For an approved access change, agree the precise effect, affected people, duration, rollback/recovery and verification with the owner before implementation. If current access is already correct, proceed to verification without an unnecessary grant. If the correct owner or business authority is uncertain, stop and resolve that first.

Review effective access through direct, group, inherited and link routes with the real owner. Correct existing access can go straight to verification; a change needs authority and a narrow scope.

Make only the approved scoped change

Use the resource’s supported control and retain an access record.

The authorised resource or role owner should use the supported administration or sharing control for that exact task. Confirm the recipient, resource and view/edit or role level before applying the agreed change. Follow the organisation’s approval process and record the owner, reason, scope, date and review or expiry requirement.

Do not broaden an entire site, expose an Anyone link, grant application consent or turn off a security requirement to fix one access denial. An app’s request to read data is a separate assessment of the app, publisher, permissions and organisation policy. It is not an ordinary user permission implied by installing the app.

When somebody changes role or leaves, arrange the organisation’s full access and data-continuity process. Removing one group or link does not establish that all access or active sessions have ended, and deleting an account can affect business data. Keep offboarding, record retention and ownership transfer within an authorised plan rather than treating this guide as a complete offboarding checklist.

Apply only the authorised resource or role change and keep its approval/review record. App consent and staff offboarding require their own scope; one removed permission does not prove complete access revocation.

Next: Verify the agreed sign-in and access scope

Handle blocked registration or access with the owner

An error needs diagnosis, not a policy bypass.

Privately record the exact error, time, account/organisation, intended resource and whether the failure is registration, authentication or resource access. Avoid screenshots containing setup codes, access tokens or private document content. Ask the responsible administrator to assess relevant registration, sign-in and policy evidence; ask the resource owner to check effective permissions.

If the only method is lost or unavailable, use the organisation’s established identity-verification and recovery process. Do not approve random prompts, borrow another identity, remove all methods, reset every device or disable MFA to continue. A report-only policy result or permitted method setting is not proof the intended enforced access works.

Unexpected approvals, changed methods or suspicious access need the organisation’s incident process promptly. A successful password reset or newly working app is not proof that an account compromise is resolved. For ordinary unresolved access, keep existing permitted work and recovery intact while the owner investigates.

Separate registration, sign-in and resource errors. Preserve approved access, use the real owner’s evidence and recovery process, and route unexpected changes through incident handling.

Verify the agreed sign-in and access scope

Use a harmless task and record what remains unchecked.

For method setup, confirm the correct work account and complete an authorised sign-in verification using the new method while retaining an approved recovery route. Do not delete the old method or sign out every session as a test. The administrator should review actual policy/sign-in evidence for the agreed users and applications; a single prompt is not a full enforcement audit.

For resource access, use a harmless approved sample in the intended location with the actual authorised participant. Confirm the required view or edit task works and have the owner review relevant effective permissions. Do not invite an unauthorised person, expose real confidential content or attempt prohibited access to demonstrate a denial.

Record the scope, results, exceptions and owner of any follow-up. Arrange the agreed review/expiry process and stop unrelated changes once the intended task works with the required safeguards. This limited check does not establish every user, application, file permission, session or offboarding outcome is correct.

Face-down smartphone, closed navy laptop and cream notebook on a pale kitchen counter

Verify the intended method and permitted resource task, retain approved recovery and record gaps. Stop unrelated changes; this sample is not a tenant-wide policy, permissions or offboarding audit.

Back to the symptom checker

Get help with account, policy or permission scope

Bring the task, identity, resource and exact error without secrets.

Use the organisation’s known IT, identity or resource owner for uncertain MFA enforcement, lost-method recovery, privileged roles, unfamiliar app consent or a wider access review. Privately provide the intended account/organisation, task, affected resource and relevant error/time. Never send passwords, enrolment QR codes, one-time codes or the full private data set.

Our Microsoft 365 support can discuss a scoped assessment and appropriate owner involvement. Agree authority, continuity, approved changes and verification before work. A working login alone does not prove every permission or policy is right, and this guide does not certify compliance or guarantee an account cannot be compromised.

Use the known owner for unresolved identity, policy or access questions. Agree the actual scope and limits, provide minimal private evidence and keep credentials and enrolment secrets out of support material.

Frequently asked questions

Does installing Microsoft Authenticator turn MFA on everywhere?

No. Installing the app, registering a method and enforcing MFA are different tasks. The administrator must check the actual tenant policy and scope. An existing session or supported method may satisfy a requirement without a fresh prompt at every sign-in.

Can I delete my old phone’s method as soon as I add a new phone?

Keep an approved working route until the replacement is enrolled and verified and the organisation’s recovery arrangements are confirmed. Then follow its authorised method-removal process. Do not remove your only usable method as a test.

Is an Authenticator push approval phishing-resistant?

An ordinary push approval or code is not automatically phishing-resistant. Supported passkey/FIDO2 methods, including permitted passkeys in Authenticator, work differently. The right choice depends on the actual device, method and organisation policy.

Should I give someone an admin role so they can open a shared file?

No. Confirm the correct identity, source and intended view/edit access with the resource owner. File permissions, group membership, licences and administrative roles are separate controls. A broad admin role is not an appropriate shortcut for ordinary document access.

Does removing a sharing link remove all access?

Not necessarily. Access may remain through direct permissions, groups or inheritance. Have the owner review effective access. Staff offboarding and incident response also involve other controls and continuity needs; one changed link is not proof that all access or sessions have ended.

More practical IT guides

View all guides

Share this guide