We would urge large organisations to be extremely cautious about enabling Frontier for all users.
Frontier is specifically about exposing users to features that are still under development so Microsoft can gather feedback. That means organisations need to expect change, incomplete functionality, numerous bugs and features that may evolve substantially or potentially never reach General Availability.
The challenge is that enterprise customers are already dealing with significant change and product issues in functionality that has reached General Availability. Extending preview functionality across an entire company introduces considerably more operational risk.
Support and documentation are another concern. With many preview capabilities, we've found that documentation is limited or unavailable, support options are constrained, and functionality can change before administrators have had an opportunity to understand, test and communicate the impact.
We've also recently discovered an important interaction between TAP and Frontier. Our TAP participation can override the Frontier configuration we were relying on. That was not something we had previously understood, and it's significant enough that we're now reviewing whether we should remain in TAP.
For enterprise administrators, the rate of change across Microsoft 365 and Copilot is becoming extraordinarily difficult to contain. When you combine frequent updates, undocumented changes and administrative settings that don't always produce the expected user experience, keeping governance, support, communications and training aligned becomes extremely challenging.
Enabling Frontier tenant-wide would multiply that problem exponentially.
Our advice to other enterprise customers: treat Frontier for what it is, an early-access (alpha release) and feedback programme. Use a controlled cohort of users who understand that they're testing evolving functionality and expand only where the benefits justify the additional operational overhead.
Frontier may be valuable, but enabling it broadly could be far more disruptive than it initially appears.