Apple stops pretending one model can do everything well
For decades, Apple built Siri as a walled garden — a voice assistant that spoke only in Apple's own tongue. Now, facing a world where AI capability has become the defining measure of a platform's worth, the company is preparing to open that garden to outside voices, letting users summon Gemini, Claude, and others from within Siri itself. The shift, expected with iOS 27 and a likely WWDC 2026 unveiling, is less a concession than a strategic reorientation — Apple trading exclusive control for a seat at the center of an expanding AI economy.
- Siri's long-standing reputation as the weakest major voice assistant has created real competitive pressure, and Apple can no longer afford to let its AI limitations define the iPhone experience.
- The plan to route user queries to rival models like Gemini and Claude marks a striking break from Apple's historically closed ecosystem philosophy.
- Apple is rebuilding Siri from the ground up — adding a standalone app, conversation history, document uploads, and text generation tools that transform it from a voice command shortcut into a full AI interface.
- A revenue-sharing model on third-party AI subscriptions gives Apple a financial incentive to open the platform, turning competitors into partners who pay for access to Apple's vast user base.
- Internal testing under the codename Campo is underway, though no plans are confirmed — WWDC 2026 is expected to be the moment Apple shows its hand.
Apple is preparing to open Siri to third-party AI models in a way the company has never attempted before. Under plans expected to arrive with iOS 27, users would be able to route their questions from within Siri to competing services like Google's Gemini or Anthropic's Claude — a meaningful departure from the closed, Apple-controlled system that has defined the assistant since its debut. Bloomberg reports that Apple is testing this capability internally under the codename Campo, with a likely preview at WWDC 2026.
The redesign goes well beyond adding new AI options. Apple is building a dedicated Siri application where users can maintain conversation history, upload documents, switch between voice and text, and organize past exchanges — functionality that resembles existing chatbot platforms far more than the ephemeral, voice-first Siri users know today. Deeper system integration is also planned, including an "Ask Siri" toggle for sending selected content into a new conversation and a "Write with Siri" option embedded in keyboard menus.
The financial logic is significant. By allowing third-party AI services to connect through Siri, Apple positions itself to take a share of subscription revenue those services generate — meaning competitors become a new income stream rather than a threat. It is a model that lets Apple sidestep its own AI limitations while keeping users anchored to the Apple ecosystem.
Siri has long trailed rivals like Google Assistant in raw capability, and this transformation appears designed to close that gap without requiring Apple to win the underlying model race itself. Whether the plans hold as described will become clear when Apple takes the stage in June.
Apple is preparing to crack open Siri in a way the company has never done before. Rather than keeping its voice assistant locked within Apple's own AI ecosystem, the company plans to let users route their questions to competing services like Google's Gemini or Anthropic's Claude—all from within Siri itself. The shift is expected to arrive with iOS 27, likely unveiled at Apple's developer conference in June 2026.
The move represents a significant departure from Apple's historical approach to Siri, which has operated as a closed system tied to Apple's own technology and, more recently, to OpenAI's ChatGPT through a partnership. Under the new model, users would gain genuine choice. Someone asking Siri a question could direct that query to whichever AI service they prefer for that particular task, rather than being funneled through a single pipeline. Bloomberg's reporting suggests Apple is actively testing this capability internally, working under the project codename Campo.
Beyond the choice itself, Apple is redesigning how Siri works across the entire device. The company is building a standalone Siri application—a dedicated space where users can access their conversation history, upload documents for analysis, switch between voice and text input, and organize their interactions. The interface is expected to resemble existing chatbot applications, with features like saved conversations, searchable chat history, and the ability to mark favorite exchanges. This represents a fundamental shift from Siri's current incarnation as a voice-first, ephemeral assistant.
The integration runs deeper than just the app itself. Apple is weaving Siri more tightly into iOS's core functions. Users will find an "Ask Siri" toggle that lets them send selected content directly into a new conversation with the assistant. A "Write with Siri" option is expected to appear in keyboard menus, allowing users to generate or edit text without leaving their current app. These changes suggest Apple sees Siri not as a voice command tool but as a full-fledged AI interface woven into daily device use.
There is a financial dimension to this openness. By allowing third-party AI services to integrate with Siri, Apple positions itself to take a cut of subscription revenue generated through these integrations. A user who signs up for Claude through Siri would generate a transaction that Apple could monetize. This revenue-sharing model may explain why Apple is willing to dilute its control over the assistant experience—the company gains a new income stream from the AI services market without building all the models itself.
The timing matters. Siri has long been considered one of Apple's weaker competitive offerings compared to rivals like Google Assistant or Amazon's Alexa. By opening the platform to best-in-class AI models, Apple can leapfrog its own limitations while maintaining the relationship with users. Users stay within the Apple ecosystem; Apple captures a portion of the value created by third-party AI companies.
Apple has not officially confirmed any of these plans. The company typically reserves major announcements for its developer conference, and Bloomberg's reporting suggests that internal plans can still shift. But the convergence of multiple reports pointing toward the same direction—third-party AI integration, a standalone Siri app, deeper system integration—suggests the company is serious about this transformation. WWDC 2026 will likely be the moment when Apple reveals how thoroughly it intends to remake Siri from a proprietary voice assistant into something closer to a platform.
Citações Notáveis
Apple is testing a standalone Siri app for devices such as the iPhone, iPad, and Mac later this year, allowing users to access their past interactions directly.— Bloomberg's Mark Gurman
A Conversa do Hearth Outra perspectiva sobre a história
Why would Apple give up control of Siri to competitors like Google and Anthropic? That seems to undermine the whole point of owning the assistant.
Apple isn't really giving up control—it's changing what it's controlling. Instead of owning the AI model, Apple owns the relationship with the user and the transaction. Every time someone uses Claude through Siri, Apple takes a cut. That's a different kind of power.
But doesn't that mean users might just go directly to Claude instead of using Siri at all?
Possibly. But the friction matters. If Claude is already integrated into Siri, and you're already in Siri, you're more likely to use it. Apple is betting that convenience and integration are worth more than exclusivity.
What about the standalone Siri app? Why does Siri need its own app when it's already built into iOS?
Because right now Siri is ephemeral—you ask it something, it answers, and the conversation disappears. A standalone app lets you build a relationship with Siri over time, save conversations, search your history. It becomes something you return to, not just something you summon.
Is this Apple admitting that its own AI models aren't good enough?
Not exactly. It's Apple recognizing that different AI models are better at different things. Claude might be better at writing, Gemini at research. By letting users choose, Apple stops pretending one model can do everything well.
When does all this actually arrive?
The preview is expected at WWDC in June 2026, with iOS 27. But Apple's plans can change, and there's always a gap between announcement and actual availability. The real question is how smoothly the integration works when it ships.