Apple Reportedly Wants a Proactive Siri, With 2027 Timing Still Tentative

Federighi's consolidation of mobile and desktop operating-system leadership puts the reported instruction within a longstanding software organization.

That organization has an uneven delivery record to work past. Apple introduced Apple Intelligence in June 2024 as an AI system for iPhone, iPad and Mac that would use personal context and take action across apps. In March 2025, Apple acknowledged in a statement reported at the time that personalized Siri features would take longer than expected, with delivery anticipated in 2026 rather than the previously indicated 2025.

Apple's original examples included finding a podcast recommended in a message or email and sending selected photos to another person. Cross-app action was already part of that plan; continuing to pursue a goal without fresh prompts is the further ambition now being described.

The stakes extend beyond a feature name. In the March 2025 report, Apple said Siri handled 1.5 billion user requests a day. That company-provided volume counts requests, rather than distinct users or successful tasks, and does not measure demand for a future agent. It shows the scale of the existing assistant Apple would be changing.

Apple's WWDC24 developer graphic illustrated app domains for integrating functionality with Siri and Apple Intelligence. Graphic: Apple. Source: Apple developer session on Siri and App Intents.

App integration offers a route, not a proven cost advantage

The accounts describe OpenAI's Dots, Meta's Muse and Grok as examples of agents that can track goals and continue pursuing them. Apple would be following a direction already taken by competitors, rather than introducing the idea of persistent assistance.

The second account also describes Muse and Grok Bot as using their own virtual machines to handle background tasks. It argues that Apple's existing frameworks, including App Intents, could reduce the need to operate software through computer-use capabilities and potentially lower the cost of some tasks. App Intents lets developers expose app functionality to Siri and Apple Intelligence, as Apple explains in its 2024 developer session.

That is a hypothesis about Apple's integration advantage, not a measured saving. The available reporting supplies no comparable task costs or performance results. Apple's earlier promise of cross-app actions helps explain the possible route, but it does not establish that Siri can sustain reliable work across a long-running goal.

Apple's infrastructure also goes beyond the device. In its original Apple Intelligence announcement, the company described on-device models alongside Private Cloud Compute, which uses larger models on servers powered by Apple-designed chips for more complex requests. That architecture does not establish how a persistent Siri agent would run or how much capacity it would require.

There are earlier commercial alternatives, too. The March 2025 report described Google adding Gemini to its assistant in 2024 and Amazon announcing an AI overhaul of Alexa in February 2025. Amazon then said the new capabilities would be included for Prime subscribers and cost $19.99 a month otherwise. Those historical terms illustrate a rival subscription approach; they do not establish current pricing or a like-for-like comparison with the proposed Siri.

Apple introduced Apple Intelligence as free for users in 2024. The new reports give no price or supported-device list for the proposed proactive assistant.

The reported milestone shifts time from integration to coding and exploration. Chart: Magica, using retained reporting citing Bloomberg's Mark Gurman. Source: 9to5Mac.

What the next development cycle must establish

Apple is also reportedly changing the first iOS 28 development cycle. The second account says the M1 milestone allocates six weeks to coding, exploration and feature development, followed by two weeks of integration, instead of the usual four weeks for each phase. The total remains eight weeks; the change shifts time toward exploration rather than shortening the cycle.

Employees reportedly welcome the extra room to develop ideas before locking them in. That feedback establishes neither faster delivery nor a link between the process change and a particular Siri release.

The next evidence needed is a defined product: which goals Siri could pursue, which apps it could act through, when it would seek permission and how reliably it could complete work. Hardware requirements, cloud capacity and pricing would determine who could use it and on what terms. The reported engineering change gives Apple more exploration time within one milestone; it leaves those product decisions unresolved.