WorkThe story, in brief

Fast AI, slow rollback: the risk facing Apple IT teams

69% of IT teams can't roll back a bad AI-generated config in an hour. Fleet's survey exposes the deployment-velocity gap that's about to bite enterprise admins.

Illustration of two anonymous hands arranging task cards around an amber tool on a shared desk.
People, judgement and the changing nature of work.AI illustration by KeyNews
The KeyNews take

Why it matters

Enterprise IT is deploying AI-generated configurations and scripts at speed, but lacks the rollback and review infrastructure to contain failures at scale. This operational gap—especially in MDM (mobile device management) where changes bypass platform security—is a growing risk as AI authoring accelerates across macOS, Windows, and Linux.

The key facts

15 to know
  1. 86% of 250+ IT practitioners allow AI-written output to reach production after review

  2. 61% used AI to author MDM profiles in past month; 62% used it for scripts/code

  3. 69% cannot roll back a bad configuration within 1 hour; 43% need >24 hours

  4. 26% can recover same day but only with manual intervention

  5. 7.6% of Fleet's customers are Mac-exclusive shops; most manage multiple platforms

  6. 36% of respondents manage Apple devices through Intune (Windows-first platform)

  7. 31% cited keeping up with Apple's yearly OS release cycle as hardest part of job

  8. Fleet CEO McNeil recommends GitOps with YAML, Git-based pull-request review, and declarative config over imperative scripts

  9. 86% of 250+ surveyed IT practitioners allow AI-written output to reach production after review

  10. 69% cannot roll back a bad configuration within one hour; 43% need more than 24 hours

  11. 26% can recover same-day only with manual intervention

  12. 31% cited keeping pace with Apple's yearly OS release cycle as hardest part of job

  13. Fleet survey; full data available via Fleet Device Management website

  14. MDM is a trusted path that can bypass macOS platform security prompts, making bad MDM-delivered changes highly consequential

  15. Suggested remediation: GitOps workflow (Git repository, pull-request review, automation, commit revert) — pattern works with any tool with an API

Go to the source

Computerworldcomputerworld.com

Publisher excerpt: There’s a big disconnect between the rate at which IT is deploying various kinds of AI-generated output and the speed with which it can roll those changes back when things go wrong, warns a new report from Fleet Device Management. It’s almost as if the rush to embrace AI has eclipsed the need to…
Read original report
Back to today's editionMore work news

Keep reading

Related stories

More from Work