Apple's iOS 27 Brings AI to Every Employee's iPhone

Apple WWDC kicks off next Monday. The headlines will focus on design language, new hardware specs, and whether the iPad Pro gets another camera bump.

The announcement worth paying attention to if you run a business is something called Extensions.

Extensions is Apple's name for a new framework in iOS 27 that allows third-party AI providers to integrate directly with Siri, Writing Tools, and Image Playground at the operating-system level. Users will be able to choose Claude or Gemini to power the writing assistance across every app on their phone. Each model gets its own Siri voice. Apple already has a deal with Google to use Gemini for certain native Siri features, and the Extensions system broadens that opening to any provider that ships an App Store integration.

This is a meaningful shift. iOS 18 added ChatGPT as a fallback for Siri questions, which felt like an experiment. iOS 27 turns that experiment into infrastructure. By this fall, every new iPhone will include a user-selected AI model powering the keyboard, the writing assistant, and the voice interface. For most of the business world, that means AI is no longer something you adopt. It's something you manage.

Think about what actually happens in a small business when this ships. Your employees have iPhones. They use them for email, for customer communication, for drafting documents, sometimes for work in apps that touch your data or your clients. In iOS 27, each of them will be choosing whether Claude or Gemini writes their replies, refines their sentences, and handles Siri requests in the context of your business. Without any policy decision on your end, the phone makes that call for you.

That's not a crisis by itself. AI in a keyboard is a long way from AI running your operations. But the pattern it sets matters.

The businesses that end up with coherent AI practices are almost never the ones who woke up one day and decided to have a policy. They're the ones who started asking the right questions before the tool was already in use. The question worth sitting with right now is not which AI is best for your employees. That's the wrong level. The question is: what do you actually want AI doing in your business, and what do you want people doing instead?

That sounds abstract, but it has a concrete version. Take one work context where your team handles sensitive information: client communication, financial data, HR matters. Write down, in a sentence, what you want to be true about how AI touches that context. You don't need a policy document. You need a position. "We don't run client data through external AI tools" is a position. "AI is fine for first drafts, but a human reviews everything before it goes out" is also a position. Not having thought about it is a position too, just not one you chose.

The Apple story is also the latest data point in a pattern that's been building for a year: AI is becoming infrastructure, the way cloud storage and email became infrastructure. Nobody asks whether employees use cloud storage. You decide what your policies are around it and you move on. The same shift is happening with AI, just faster and with more variation in what the tools actually do.

WWDC will tell us how much flexibility users actually get, which providers Apple certifies, and what the privacy implications look like in practice. Those details matter, and the coverage next week will be worth reading. What's already clear is the direction: Apple is treating AI as a layer in the operating system, not a feature inside an app. That's a decision about permanence.

The businesses paying attention to this aren't trying to control whether AI enters their workplace. That ship has sailed. They're trying to be thoughtful about how, so that when iOS 27 ships in the fall, they've already had the conversation that most companies will scramble to have afterward.