Apple is reimagining Siri for iOS 27. It is a bigger deal than we might expect. Bloomberg sketches reveal what’s coming. Later this year we get the real thing. The illustrations show a pivot from reactive tool to agentic helper.
Siri will shift from single tasks to complex workflows. No step-by-step instructions needed.
Timelines are tight. WWDC happens June 8 to 12. iOS 27 drops in mid-September. Gurman bases his visuals on insider info. Names stay hidden because the software isn’t public yet. An Apple rep stayed silent on this round of questions.
Where does it live?
Forget the floating orb. Siri lives in the Dynamic Island now. It’s always there at the top. Pill-shaped. Always on.
Two ways to wake it. The old school way still works. Say “Hey Siri”. Hold the side button. You get the animation. It’s fine for voice search.
The new way feels different. Swipe down from the top center. You hit the Search or Ask interface. Type or speak. Results pop up in text cards. Swipe down again and you chat. It feels native. It feels integrated.
You won’t just talk to Siri. A menu lets you pick agents. ChatGPT. Gemini. Anthropic’s Claude. Apple has ties to OpenAI. They tested Gemini and Claude integration too. The same panel shows weather and suggestions. Eight apps usually sit there. Familiar ground amidst change.
Don’t worry about your notifications. The center is still there. Just move your hand. Swipe from the top left now. The top center swipe goes to Siri. That’s the shift.
New apps and eyes
Siri gets an actual app. Finally. Like ChatGPT has one. Like Gemini. It keeps history. You can look back at past chats. Voice or text works either way.
Upload docs. Upload photos. Let Siri analyze them.
It reads the room. News queries pull rich cards. Sports scores show up fast. It digs deeper into your data. Text messages. Emails. Calendar events. Siri drafts replies. It checks your schedule. Tells you if you’re booked or free.
The camera joins in too. Scan an object. Let an AI agent analyze it. Google Search runs the query. Similar to how Google Lens works. It sees what you see.
The cloud question
The Information reports another layer. Privacy is the angle. iPhones process simple AI queries locally. The chips in Macs, iPhones, and Watches do the heavy lifting. No server trip needed for basic stuff.
Complex requests go elsewhere. Specifically to Google.
Why not build giant data centers? It’s expensive. Outsourcing hard tasks saves cash. Google’s Gemini trains smaller models for Apple devices. A symbiotic arrangement. One question lingers though. Can you trust this balance enough to keep your secrets safe?
