Apple is pushing to rebuild Siri into something closer to an autonomous AI agent — a system that tracks a goal and acts on it without being walked through each step — according to Bloomberg's Mark Gurman, whose reporting describes internal pressure from Apple's leadership to move faster on the project.
Gurman reports that Apple CEO Tim Cook's lieutenant, hardware engineering chief John Ternus, told software chief Craig Federighi in a closed-door meeting that the company needs to "accelerate" its roadmap and move "beyond merely making necessary improvements to Siri." The overhaul could arrive as early as 2027, though Apple has made no official announcement and has not confirmed a timeline or feature set.
Catching up to a moving target
The urgency reflects how quickly the competitive landscape has shifted. Rivals including OpenAI, with its Dots assistant, Meta's Muse, and xAI's Grok have already shipped always-on agents capable of taking multi-step actions with minimal prompting, leaving Siri — which has lagged on generative AI capability since Apple Intelligence's rocky 2024 rollout — looking increasingly dated by comparison.
Nearer-term, Gurman reports that iOS 27.3 will give Siri finer-grained control over Apple's own apps, with support for third-party app integration potentially following in the spring. Separately, Mike Rockwell, the executive who has led Siri's AI development, is said to be weighing whether to step back from the role — a detail that, if accurate, would add leadership uncertainty to an already pressured timeline.
What's actually confirmed
None of this is official: Apple has not announced an agentic Siri, a release date, or specific capabilities, and the reporting should be read as describing internal deliberation rather than a settled product plan. Apple has a history of letting AI features slip — Siri's personalized-context upgrade was delayed multiple times during 2025 — so any 2027 timeline carries real risk of further slippage.
If Apple does ship a more autonomous Siri, the bigger question is whether it can match the reliability rivals have already demonstrated in production, given the company's uneven recent track record on generative AI features.
