Apple has quietly posted job listings that lay out its plans for a full autonomous AI agent — one that browses the web, works across apps, runs commands, writes and debugs code, and controls devices on the user's behalf. Bloomberg reported the listings on Thursday, noting that one posting seeks a senior engineering manager for a "proactive, agentic AI assistant," and trade outlets from The Verge to MacRumors and 9to5Mac quickly picked up the signal: Apple is formally entering the agent race that OpenAI, Anthropic, and Meta have been sprinting through all year.

The postings arrive just days after Bloomberg's Mark Gurman reported that Apple wants to remake Siri into an autonomous agent that proactively carries out tasks without users initiating every step — with the capability potentially arriving as soon as next year. Together, the two reports sketch the clearest picture yet of where Apple's long-running AI rebuild is headed. For a sense of how fast the rest of the industry is moving in the meantime, see our roundup of the latest AI developments.

What the job listing actually says

The most revealing posting is for a "Senior Engineering Manager, Agentic AI – Proactive" role. Its summary, quoted by 9to5Mac, reads less like a job ad and more like a product manifesto:

> "Imagine an AI agent that can use a computer the way a capable person does. It browses the web, works across apps, runs commands, writes and debugs code, and calls services, and it chooses the right tool for each step. It does all this privately and with the polish people expect from Apple. Our team is building the foundation for that future, and we're looking for a senior engineering leader to make it real."

The listing goes further, describing agents that can "plan and carry out multi-step tasks across apps, browsers, command lines, code, and APIs, and they can coordinate specialized sub-agents along the way" — language that maps almost one-to-one onto what frontier labs describe when shipping their own computer-use agents. As 9to5Mac notes, the wording suggests development is still at an early stage; Apple is hiring the leadership to build the foundation, not advertising a finished product.

The Siri connection — and what "proactive" means

The new listings follow Gurman's report from days earlier that Apple intends to make Siri behave more like an autonomous agent: one that acts on intent, chains together multi-step work, and checks in only when it needs permission or judgment. Per that report, the company could introduce such capabilities as soon as next year.

"Proactive" is the operative word. Today's assistants — including the overhauled Siri Apple has been rolling out after repeated delays — mostly respond to commands. An agent that initiates work, monitors context, and coordinates across devices is a different product category, and Apple's own job ad draws the line explicitly: the system should use a computer "the way a capable person does," choosing the right tool for each step rather than waiting to be told which tool to use.

Why Apple feels the heat

The competitive backdrop explains the timing. Meta's Muse assistant has become one of the fastest-growing apps in US history — 10 million downloads in 40 days, according to Sensor Tower data cited by 9to5Mac — and as of Thursday sat at No. 2 among free US App Store apps, behind only TikTok. OpenAI and Anthropic ship their own always-on and proactive agents, and 9to5Mac attributes much of the surge in agent interest to the breakout success of OpenClaw, the autonomous agent that made computer control a mainstream expectation this year.

Apple has watched from the sidelines while also raising the walls. Last month it announced tighter macOS privacy controls around Full Disk Access, with changes in the macOS 27.3 beta giving users granular control over which apps and data third-party AI agents can touch — a move widely read as both a privacy safeguard and a way to make sure Apple-defined agents get the best integration on Apple hardware.

Privacy as the differentiator

That word "privately" in Apple's listing is doing a lot of work. Apple cannot out-scale the frontier labs on raw model capability overnight, but it can position the agent as the one that sees your screen, email, files, and messages without shipping that data off-device. The pairing of a first-party autonomous agent with stricter permissions for third-party ones gives Apple a coherent story: the most capable computer-use agent on a Mac should be the one Apple built.

There is also a platform angle. An agent that runs commands, calls services, and coordinates sub-agents becomes a new layer between users and apps — the kind of layer that has historically reordered software economics on whichever platform controls it. Apple's cautious, privacy-forward implementation is a bid to own that layer on its own terms before third-party agents cement it elsewhere.

What to watch next

Three things to watch as this develops. First, more hiring: a senior manager posting usually precedes a wave of engineer listings, and additional agentic roles on Apple's careers site would confirm the team is scaling. Second, developer frameworks: if Apple plans to let third-party apps plug into a system-level agent, that will surface in OS betas long before marketing does. Third, timing slippage: Apple's Siri overhaul arrived late, and 9to5Mac's analysis notes the agent landscape could look very different by the time Apple ships — the industry moved toward autonomous agents while Apple was still finishing its assistant-era roadmap.

For now, the concrete facts are the listings themselves: Apple is hiring to build an autonomous, multi-step, privacy-focused agent that browses, codes, and controls apps — and it has told the world about it in its own job ads.

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →