
Ales Nesetril on Unsplash
A new analyst report, backed up by recently posted job listings, makes clear that the company is now actively building its answer to that challenge.
The pressure on Apple to match the agentic AI capabilities already shipping from its competitors has been building for months.
Bloomberg reported earlier today that Apple recently posted new job listings that offer a detailed look at its plans to develop autonomous AI agents capable of performing complex multi-step tasks on a user's behalf. The listings, surfaced and detailed by 9to5Mac, describe a system that would browse the web, write and debug code, control apps, run commands, and call services, all without requiring the user to initiate each step.
One of the most notable listings is for a Senior Engineering Manager, Agentic AI, Proactive role, which carries a job summary that leaves little to the imagination about the scope of what Apple is building.
"Imagine an AI agent that can use a computer the way a capable person does," the listing reads. "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."
The listings also describe plans for agents that can coordinate specialized sub-agents to complete broader goals, allowing the AI system to break a complex task into components and dispatch the right tool for each. That architecture mirrors what frontier AI labs, including Anthropic, OpenAI, and Google, have already built into their own agentic products.
The timing of these job listings is not coincidental. The recent Bloomberg report arrives just eleven days after a separate Bloomberg report from Mark Gurman described Apple's interest in turning Siri AI into something far more proactive, with the ability to monitor ongoing situations and take action without being explicitly prompted each time (via 9to5Mac).
Taken together, the two reports suggest a connected strategy rather than parallel, unrelated efforts.
The focus of the job listings is on Mac-based agents rather than iPhone-based ones. That distinction matters for practical reasons: Mac users are the audience most likely to adopt agentic workflows for tasks like software development, research, file management, and long-running productivity work. Mobile AI agents present additional constraints around power, interface, and the kind of complex multi-step tasks that benefit most from an autonomous agent approach.
The focus on Mac-first development is also consistent with Apple's existing macOS privacy controls. In November 2026, Apple introduced new privacy controls in the macOS 27.3 beta that give users more granular control over Full Disk Access, specifically to let people decide which apps and AI agents can see which files and folders.
Those controls create the guardrails within which an Apple-built agent would need to operate, and building that agent first-party gives Apple direct control over how it interacts with those privacy boundaries.
Bloomberg's Gurman reported that Apple could introduce autonomous AI agent capabilities as early as 2027, with the agentic Siri work seen as part of the broader iOS 28 and macOS 28 development cycle that Apple typically previews at WWDC in June. Apple has not confirmed any timeline or feature set for the agent work described in the job listings, and the listings themselves note that development is still in an early stage.
The competitive pressure Apple is responding to is significant. Meta's Muse reached 10 million US downloads in just 40 days after its September 2026 launch, surging to the top of the App Store charts and sitting at second place among free apps as of this writing. OpenAI's ChatGPT Dots is operating as an always-on background agent for subscribers. Anthropic's Claude Code and its broader agentic products are widely used by developers for multi-step programming tasks.
Meanwhile, Apple only launched its long-awaited Siri AI overhaul in iOS 27, arriving in a market where the competitive bar had already moved substantially further ahead.
The language of the Apple job listings, while detailed, does not indicate that shipping is imminent. Building an agent that operates across apps, browsers, the command line, and external APIs while doing so "privately and with the polish people expect from Apple" is a substantial engineering and design challenge, and Apple has historically been willing to take longer to ship something it considers ready rather than arriving first with a rougher product.
Autonomous AI agents are software systems powered by artificial intelligence that can carry out multi-step tasks on a user's behalf with minimal or no manual input after an initial instruction. Unlike a standard AI assistant that responds to individual prompts and waits for the next one, an autonomous agent can be given a goal and then independently plan the steps needed to achieve it, taking actions such as browsing the web, running code, sending emails, filling out forms, or controlling other apps along the way.
The defining characteristic is autonomous action across multiple platforms and over an extended period of time, with the agent making its own decisions about which tools to use at each step. Examples already on the market include OpenAI's ChatGPT Dots, Meta Muse, Anthropic's Claude Code, and Google's Project Mariner, all of which operate in browser or desktop environments with varying degrees of autonomy.
Based on the Bloomberg and 9to5Mac reporting, Apple's autonomous agent development is closely linked to Siri AI and the broader Apple Intelligence platform. The earlier Gurman report explicitly framed the agentic Siri work as a Siri AI evolution, and the job listing for the Senior Engineering Manager role sits under the same Siri and Apple Intelligence organizational umbrella.
It is likely that any agent Apple releases would be positioned as a Siri AI capability, extending the platform's existing abilities to take actions across apps and perform multi-step tasks into something more proactive and long-horizon. Apple has not confirmed the branding or the precise relationship between its agent work and Apple Intelligence, so the exact product form remains speculative until an official announcement is made.
The focus on Mac-based agentic AI rather than iPhone-based agents comes down to what autonomous agents are most useful for and where the technical environment makes them most feasible. The tasks described in Apple's job listings, browsing the web, writing and debugging code, running command-line commands, calling APIs, and controlling apps, are tasks that Mac users perform in professional and developer workflows rather than in mobile-first consumer contexts.
Mac also provides a more flexible operating environment for agent actions, with access to a file system, a terminal, full web browsers, and the ability to run long-horizon tasks without the battery and interface constraints that make extended autonomous operation on a phone more complex. Additionally, the macOS privacy controls Apple is currently tightening in macOS 27.3, specifically around Full Disk Access and app-level permissions, suggest that the security groundwork for a Mac-first agent is being laid intentionally.
