Apple Explores Turning Siri Into an Agent That Completes Tasks for You
Apple is exploring a version of Siri that works as an agent, carrying out multi-step tasks on the user's behalf rather than only answering questions or running single commands, according to a Bloomberg Power On newsletter published November 8, 2026. The newsletter's headline was “Apple Explores Turning Siri AI Into ChatGPT Dots-Like Agent.” The reporting describes an effort to move Siri from a command-and-response assistant toward software that completes multi-step work.
According to The Verge, Bloomberg's Mark Gurman reported that John Ternus, whom The Verge identifies as CEO, is pressing Craig Federighi, whom it describes as AI czar, to accelerate the Siri roadmap. The accelerated plan aims to deliver an AI agent that competes with OpenAI's Dots and Meta's Muse. The Verge also states that Federighi has already delivered a functional version of Siri AI on an incredibly short timeline.
In the Power On discussion, Ternus said Apple now has “the kind of Siri that we always wanted” and called the offering a “new chapter in AI.” Bloomberg's newsletter carried those remarks.
The rollout is staged. The Verge reports that finer-grained Siri control over Apple's first-party apps, such as rotating and cropping images in Photos and organizing Safari tabs, is expected with iOS 27.3. Integration with third-party apps is likely to arrive in the spring, according to Gurman, enabling tasks such as placing a Grubhub order or booking a table through OpenTable.
Earlier reporting sets the organizational backdrop. In January, Bloomberg reported that Apple had put John Ternus, then a CEO candidate, in charge of design. The same report described Apple's shift to Gemini as a fundamental reset of its AI strategy under Federighi, who had effectively assumed control of the company's AI. Apple's leadership page now lists Ternus as CEO and Federighi as Senior Vice President, Software Engineering.
Bloomberg News also reported in September, as carried by Reuters, that Ternus is weighing a broad overhaul of Apple aimed at sharpening its engineering focus and introducing products more frequently. The Siri acceleration fits that description of a company trying to ship faster.
The broader context here is a change in what Siri is expected to do. A voice assistant that answers questions and sets timers is a different product from one that works inside third-party apps, reads their state, and completes a transaction. An agent is software that takes actions toward a goal, not just produces text, so the language model is only part of the work. Permissions, app-level intents (the defined actions an app exposes to other software), error recovery when a restaurant or booking service changes its interface, and user confirmation all sit between the model and a finished order. The spring timeline for third-party integration will test those mechanics in public.
In my view, the most useful measure of this effort will be narrow and concrete: whether a routine task such as a dinner order or a reservation completes reliably, with the user remaining in control of each step. If that works, the practical gain is modest but real, with chores that now require switching between apps and forms becoming a single spoken request. The risks are equally concrete, including misdirected actions and the expanded set of data an agent must touch to be useful. Both belong in the assessment of what Apple ships.
