Nuage Solutions All articles
Microsoft 365

A Calmer Route to Passkeys: Phasing Passwordless Access Across a UK Estate

Nuage Solutions
A Calmer Route to Passkeys: Phasing Passwordless Access Across a UK Estate

Passkeys are no longer a curiosity confined to consumer apps. Microsoft 365 and Google Workspace both support them, and the UK National Cyber Security Centre has been clear that phishing-resistant authentication should be part of the journey away from passwords. For a small or mid-sized UK business, however, the practical question is not whether passkeys are a good idea. It is how to introduce them without breaking the applications that still expect a password, the shared device in the warehouse, or the contractor who arrives for three weeks with a personal phone.

The answer is a controlled programme, not a company-wide switch. Passkeys change the recovery model as much as the sign-in experience, so the rollout has to be designed around support, identity proofing and evidence.

Why a passkey programme is an identity change, not a setting

A password can be reset by an administrator. A passkey is a cryptographic credential held by a user and, depending on platform and policy, either synced across their devices or bound to a specific device. That distinction matters. If a laptop is lost, the recovery path is not simply a password reset. If a passkey is synced through a personal Apple or Google account, the organisation may have less visibility than it expects. If a device is shared, a synced passkey may be the wrong instrument altogether.

Decision-makers should therefore treat passkeys as part of identity governance. The rollout needs an owner, a support process, a recovery policy and a way to prove that access is becoming stronger rather than merely different.

Start with a narrow pilot, not the whole company

A pilot of 15 to 30 users is usually enough to expose the awkward cases without creating a queue at the service desk. Choose a group that uses Microsoft 365 or Google Workspace daily, works on modern managed devices, and does not rely on a shared terminal. Include at least one person who travels, one who uses a mobile device heavily, and one who is likely to lose a device. Exclude, for the first phase, anyone whose role depends on a legacy line-of-business application that has no modern authentication path.

Run the pilot for four weeks and define success before it begins. Useful measures include successful enrolment rate, sign-in completion rate, helpdesk contacts per 100 users, and the number of recovery events. A pilot that produces beautiful enrolment figures but a wave of locked-out users is not a success; it is a warning.

Decide where passwords must remain

A passkey-first policy is not the same as a password-free estate. Passwords will remain for three categories: applications that cannot support modern federation, service accounts and other non-human identities, and break-glass administrative access. The first category is the one that shapes the timeline. Do not set a date for removing passwords until the applications that require them have been retired, federated or isolated behind a gateway.

For legacy applications, keep password sign-in but protect it with conditional access, strong multi-factor authentication and network restrictions where appropriate. Record the exception, the owner and the review date. For service accounts, use managed identities or secrets management rather than a passkey designed for a human. For break-glass accounts, keep a password but store it under physical or vaulted control and test the recovery process.

The question of how long passwords remain is therefore answered application by application, not by a single corporate deadline. A sensible 2026 position is to make passkeys the default for modern cloud services while maintaining a documented password exception register for everything else.

Build the helpdesk script before the first enrolment

The service desk will determine whether the rollout feels calm or chaotic. Write the script before the pilot starts, not after the first lost phone. It should cover enrolment, a lost or replaced device, a synced passkey that has not appeared on a new device, a device-bound passkey that cannot be transferred, and a user who has forgotten which device holds the credential.

The script must also specify identity verification. Remote screen sharing is not sufficient proof of identity for a credential recovery. Use a callback to a known number, a video call with photo identification, or an in-person check depending on the sensitivity of the account. Record who approved the recovery, when, and on what evidence. That record is both a security control and an audit artefact.

Recovery policy: the part auditors will ask about

A workable recovery policy has two routes. The first is self-service, where the user still has a strong authenticated session or a second enrolled device. The second is assisted recovery, where the service desk verifies identity and issues a temporary enrolment path. Both should be logged. Neither should rely on SMS as the only factor.

Contractors need a separate line in the policy. If they use a personal device, do not enrol a corporate passkey without a clear consent and offboarding process. A temporary account with a device-bound credential or a hardware security key is often cleaner. On shared devices, avoid synced passkeys; use device-bound credentials or a kiosk account with session controls, and make sure the next user cannot inherit the previous session.

Audit evidence and the measures that matter

Passkey rollouts produce plenty of activity but not all of it is evidence. A monthly report should show enrolment by department, sign-in success rates, recovery events, helpdesk tickets per 100 users, and the number of active password sign-ins to cloud services. Trend matters more than a single figure. A falling password sign-in count alongside stable helpdesk volume is a good sign; a rising recovery count may indicate that the support script is too permissive or the user guidance is unclear.

Keep the pilot report, the recovery policy, the helpdesk script, conditional access change records and a sample of recovery approvals. These are the artefacts that demonstrate control to an auditor, insurer or client. They also help the next phase run faster.

A phased path for the 2026-27 planning cycle

In the first 90 days, appoint an owner, document the application estate, select the pilot group and write the support script. In the second phase, extend passkeys to the broader workforce in teams, keeping the exception register live. In the third phase, tackle contractors, shared devices and legacy applications. Only then should the organisation consider removing password-based sign-in for any remaining modern service.

Passkeys are a genuine improvement, but they are not a single project with a finish line. They are a change in how identity is issued, recovered and proved. UK SMEs that phase the work carefully will get the phishing resistance and the lower friction. Those that rush it will discover that the hardest part is not enrolment; it is the Monday morning when someone cannot get in.

All Articles

Related Articles

Permission Debt Comes Due When the Assistant Arrives: A UK Microsoft 365 Pre-AI Review

Permission Debt Comes Due When the Assistant Arrives: A UK Microsoft 365 Pre-AI Review

Seven Silent Leaks in a UK Microsoft 365 Estate — and the Monthly Routine That Plugs Them

Seven Silent Leaks in a UK Microsoft 365 Estate — and the Monthly Routine That Plugs Them

Cover Before the Vacancy: A Key-Person Audit for UK Cloud Estates

Cover Before the Vacancy: A Key-Person Audit for UK Cloud Estates