To roll out passkeys and MFA for business web applications, start by inventorying which apps and user groups need stronger sign-in, pilot passkeys with a small group, keep a recovery path that is no weaker than the passkey, then move users in phases while measuring help-desk load and sign-in failures.
Key takeaways
- Passkeys are phishing-resistant sign-in credentials, so they should be the default for applications that hold business or personal data.
- A rollout works best in phases that start with a pilot group, widen to all staff, and only then retire passwords for the apps that support passkeys.
- Recovery is the hardest part of the design, and a recovery route weaker than the passkey undoes the security gain.
- Legacy applications that cannot accept passkeys should keep MFA until they are replaced or connected through a supported identity provider.
What are passkeys, and how do they change MFA?
Passkeys replace a password and a second factor with a single cryptographic credential. Each passkey is a key pair: the private key stays on the user's device or in a synced password manager, and the server stores only the public key. The sign-in is tied to the exact website origin, so a look-alike domain cannot use it.
The FIDO Alliance describes passkeys as phishing-resistant credentials built on public-key standards, and its passkeys guidance explains the terminology used in the rest of this article. That phishing resistance is the main reason to adopt them, not convenience alone.
Traditional MFA adds a one-time code or push approval on top of a password. It reduces risk, but users can still type a password into a fake page and relay a code in real time. A passkey removes the password and the relayable code, which is why it fits the ordinary MFA question of 'what if the login page is fake?'.
Which applications should you move first?
Move first the applications that hold the most sensitive data, that administrators use, or that are reachable from the internet. Internal tools that sit behind a VPN come later, because their exposure is lower and their legacy integrations are often more fragile.
Build a simple inventory before any configuration work. For each application, record the login method, whether it supports passkeys or WebAuthn, who uses it, whether it has a separate admin login, and which other systems depend on its sign-in. The dependency column is often the one that surprises teams: a reporting tool that pulls data through a service account may break if the interactive login changes.
Some applications will not support passkeys at all. For these, keep MFA in place and plan an integration or replacement, rather than forcing a workaround that weakens the control. Our integration services cover connecting identity providers to applications that lack native support.
What does a phased rollout look like?
A phased rollout starts with a pilot, widens to all staff, and only then removes passwords where the policy allows it. Skipping the pilot is the most common cause of avoidable help-desk tickets, because edge cases such as shared devices and kiosk logins appear only under real use.
Phase 1: Pilot with a small group
Choose a group of volunteers who use a mix of devices and browsers. Enrol their passkeys on the primary application, then observe failed sign-ins, enrolment time and any support requests for a few weeks before expanding.
Phase 2: Enrol all users on priority applications
Make passkey enrolment part of onboarding for new staff and a scheduled task for existing staff. Keep the old password active during this phase so that users are not locked out while they adopt the new method.
Phase 3: Retire passwords where the policy allows
Only remove password sign-in once passkey use is stable and recovery has been tested. Removing passwords earlier creates a single point of failure that the help desk must handle manually.
Which fallback methods should you keep, and where do they fail?
Keep a fallback that is at least as strong as the passkey it replaces, and make it difficult to use for an attacker who only has a password. Authentication guidance from NIST, in its Digital Identity Guidelines, SP 800-63B, sets out how authenticator types differ in assurance, and it is a useful reference when you compare fallback options.
Each common fallback has a known failure mode. SMS codes can be intercepted or redirected through SIM swapping. Email reset links are only as strong as the mailbox they reach. Help-desk resets fail when staff are under pressure and verify identity by asking questions that are publicly known. Backup codes fail when they are stored next to the password.
- Device-bound passkeys on a second registered device: strong, but the user must keep at least one enrolled device.
- Synced passkeys in a managed password manager: convenient, but recovery depends on the account used to restore the manager.
- Recovery codes stored offline and issued once: good for rare cases, but they fail if they are lost or kept in the same place as the password.
- Verified help-desk reset using an approved process: workable at small scale, but it fails under time pressure unless the steps are written down and audited.
Does GDPR apply to authentication data?
Yes, where the accounts belong to identifiable people, authentication data can be personal data, and the GDPR requires appropriate security. Article 32 of Regulation (EU) 2016/679, the General Data Protection Regulation, requires controllers and processors to implement appropriate technical and organisational measures for a level of security appropriate to the risk. This article is general information and not legal or tax advice; check the obligations with your own qualified advisers. Rui Codex offers technical implementation support for identity rollouts.
Keep the data flow simple when you document the rollout. Record which identity provider stores public keys, where recovery data sits, how long audit logs are retained and who can change authentication settings. These records help you show what the system does, even though they do not replace a legal review.
Hypothetical examples of a passkey rollout
Here are two hypothetical scenarios:
In the first scenario, a team runs a customer portal and an internal ordering tool. It pilots passkeys on the internal tool because its users are few and close at hand. The pilot shows that shared warehouse terminals break the enrolment flow, so the team moves those logins to a managed kiosk process before the wider rollout.
In the second scenario, a team relies on a legacy reporting application that accepts only a password and an authenticator app code. It keeps MFA on that application, restricts access to a named group, and schedules its replacement. The team does not remove MFA from the legacy tool, because removing it would lower protection in the one place that has not yet changed.
What should a rollout checklist include?
A useful checklist covers scope, recovery, communication and measurement. Each item should have an owner and a date, so that the rollout does not stall once the pilot ends.
- Inventory of applications with sign-in method, admin access and dependencies recorded.
- A tested recovery route for each enrolment method, with the help-desk steps written down.
- A pilot group with a defined observation period and a list of success criteria agreed beforehand.
- User guidance that explains enrolment on each device type the organisation uses.
- Logs of enrolment, failed sign-ins and recovery requests, reviewed on a fixed schedule.
- A decision record for each application that keeps MFA, and the date it will be reviewed.
Measure the rollout by the same signals you used to plan it. Watch for rising help-desk volume, repeated failed enrolments and any growth in recovery requests, and adjust the plan when a signal moves. Our security and GDPR guides cover related controls such as logging and access review.
Rollouts that touch shared infrastructure should also be planned alongside your hosting controls. The post on securing cloud infrastructure describes a baseline that complements the identity work described here.
Conclusion
A passkey and MFA rollout succeeds when the design is staged: inventory first, pilot second, recovery tested before passwords are removed, and fallbacks chosen for the failure they are most likely to suffer. Legacy applications that cannot accept passkeys keep MFA until they are replaced or connected through a supported identity provider. If you want a second view on your rollout sequence or your integration path, request a project consultation at https://ruicodex.com/Contact or email info@ruicodex.com to discuss your project.
Frequently Asked Questions
Do passkeys replace MFA entirely?
Passkeys can replace a password and a second factor in one step on supported sites, because the credential proves possession of a device and a local unlock. They still need a recovery path, and some legacy applications accept only traditional MFA, so many businesses run both for a period.
Which applications should go first?
Start with applications that hold sensitive data, are used by administrators, or are reachable from the internet. Then add internal tools. Confirm that your identity provider supports passkeys for each application before you schedule the change.
What happens if an employee loses a device?
Losing a device-bound passkey should trigger a verified recovery process, not a password reset by email alone. Synced passkeys may survive device loss if the password manager account can be recovered. Decide the recovery design before the rollout begins, not after the first lost device.
Who should review a passkey rollout for compliance?
A rollout that touches employee or customer accounts should be checked against your own data protection obligations with qualified advisers. Rui Codex offers technical implementation support for identity rollouts, and the decisions about obligations remain with your organisation.
Sources
- Passkeys - FIDO Alliance
- NIST SP 800-63B Digital Identity Guidelines: Authentication and Lifecycle Management - National Institute of Standards and Technology
- Regulation (EU) 2016/679 (General Data Protection Regulation) - European Union, EUR-Lex