Inside Apple’s Biggest Siri Redesign Yet: What to Expect at WWDC 2026

The app introduces a proper chatbot-style interface that resembles Apple's iMessage, according to reports . Users will be able to maintain ongoing, back-and-forth conversations rather than issuing single voice commands. The interface also supports a complete conversation history, allowing users to search through past chats, pin favorites, and view threads in a list or grid layout . Beyond text, users will reportedly be able to add photos, images, and documents directly into Siri conversations . User-controlled retention policies, ranging from 30 days to indefinite, are also expected .

Dynamic Island Integration and a New Visual Design

The new Siri is deeply integrated with the iPhone’s Dynamic Island. When activated by the wake word or side button, a pill-shaped animation expands in the Dynamic Island with a glowing cursor and a prompt that reads "Search or Ask" . This marks a shift beyond voice-only interaction, positioning Siri as a flexible entry point for on-device search and AI-powered queries.

When a question is asked, results appear in a translucent panel. Swiping down on this card expands into a full conversation mode, allowing the chat-like experience directly from the system interface . The redesign also adopts a dark color scheme that matches the visual aesthetic of Apple's WWDC 2026 teaser graphics, with a glowing border that's particularly visible in dark mode .

Third-Party AI Model Support with the Extensions Framework

A significant shift in Apple's walled-garden approach, iOS 27 will reportedly introduce an "Extensions" framework that allows users to choose their preferred AI engine at the system level . Named partners include Google Gemini, OpenAI ChatGPT, and Anthropic Claude, which could power not only Siri but also Writing Tools and Image Playground . This is described as a first for a mobile operating system, giving users the ability to select a default AI model .

The Google Gemini Foundation

To power the core of the new Siri, Apple has signed a cooperation agreement with Google, making Gemini the foundational AI model for the assistant . Reports describe the deal as a multi-year partnership worth approximately $1 billion per year . Some reporting further specifies the use of a custom 1.2-trillion-parameter Gemini model , though Apple has not confirmed these technical details.

Apple frames the partnership through a privacy lens, stating that Gemini-powered processing will occur inside its own Private Cloud Compute infrastructure rather than directly on Google's servers . The reported architecture suggests on-device processing for basic tasks using Apple's own models, with more complex queries routed to Apple's encrypted cloud servers that run the Gemini model via an API . This ensures Google cannot access personal user data during processing according to statements from both companies . It remains to be seen how Apple will implement and publicly explain user control over these third-party models alongside the Gemini foundation .

A Note on Sources: All details described here are based on previews and pre-WWDC reports from sources like Bloomberg's Mark Gurman, MacRumors, and others . The full feature set, pricing models, and final architecture will only be confirmed when Apple officially presents iOS 27 and the new Siri on stage at WWDC 2026 on June 8 .