Two ways to reorganize around AI

Apple's WWDC framing today is explicitly 'we are not reorganizing the company around AI.' Microsoft's framing last week was 'the system around the AI is the company.' Both can be right; only one is a strategy.

  • #ai
  • #agents
  • #copilot
  • #devex
Two abstract structures in counterpoint: a loose feature cluster and a disciplined architectural lattice

Two corporate postures landed within a week of each other, and they are useful to read side by side.

Apple is taking another swing at its AI plans today, and the NYT framing is unusually direct: “unlike some rivals, it is not reorganizing around the technology.” The bet is that AI is a feature to be threaded into a hardware-and-services portfolio that already works, not a force that should reshape the company itself.

Microsoft, in its agent platform post from Build, took the opposite stance. The whole pitch is that AI alone won’t change a business; the system running it will, and that means treating agents as first-class production software with identity, governance, observability, and a lifecycle. Translated: reorganize around the system, not the model.

Two different bets. Both rational.

The Apple bet works if the platform’s distribution moat is wide enough that “good enough AI, deeply integrated” beats “great AI, loosely integrated.” Historically, that bet has paid for them.

The Microsoft bet works if enterprises actually want to operate their AI as a governed system, with reproducible behavior and audit trails, the same way they operate the rest of their production software. Most regulated-industry customers I’ve talked to over the last year have signaled, in different words, that they do.

The interesting thing for anyone advising an enterprise is that the choice between these postures is not theoretical. Inside your own org you have to pick one. Either AI is a feature you sprinkle into existing products and processes, or it’s a discipline you build a system around.

Sprinkling is faster, cheaper, and produces demos. The system approach is slower and more expensive, and produces durable production work that survives the second auditor.

If you’re an SI or ISV partner deciding where to invest, the question is not “which vendor’s stack.” It’s “which customer posture am I building for.” The accelerators, templates, and governance packages you ship look very different in each world. Picking one and going deep beats hedging.

All notes · RSS