Apple Job Listing Reveals Plans for AI Agent That Can Run Your Computer
An Apple job post spotted by Bloomberg describes an AI agent that browses the web, controls apps and writes code, with a launch possible by late 2027.
Apple is developing its own autonomous artificial intelligence agent, a system designed to operate a computer on a user’s behalf by browsing the web, moving between apps, running commands and writing software, according to a newly posted job listing first reported by Bloomberg News on Thursday. The posting, which seeks a senior manager to lead the effort, offers the clearest public look yet at how the iPhone maker plans to compete in one of the fastest-moving corners of the AI industry.
What the listing says
The role, titled “Senior Engineering Manager, Agentic AI – Proactive,” was posted on Apple’s careers site, according to 9to5Mac and the tech news site BigGo, which both cited the listing. “Agentic” is the industry’s term for AI systems that do not simply answer questions but carry out tasks, taking a series of steps with limited human supervision.
The listing, as quoted by 9to5Mac, opens by asking readers to imagine “an AI agent that can use a computer the way a capable person does.” It goes on to describe a tool that browses the web, works across apps, runs commands, writes and debugs code and calls online services, picking the right tool for each step. Apple added that the system would do this “privately and with the polish people expect from Apple.”
According to the post, Apple’s teams are building agents that can plan and carry out multi-step tasks across apps, browsers, command lines, code and application programming interfaces, and that can coordinate specialized “sub-agents” along the way. 9to5Mac noted that this design closely resembles products already offered by leading AI labs.
The news site Globe TV reported that the job spans Apple’s Siri and user-interface frameworks teams as well as its Apple Foundation Models group, which builds the company’s in-house AI models, and that the listing describes safety and privacy as “central to the design.” By Friday, the listing’s page on Apple’s careers site displayed a message saying the role does not exist or is no longer available. The reports did not include any comment from Apple.
Timing and the Siri connection
Bloomberg reported that Apple has considered launching such an agent as soon as the end of next year, meaning late 2027, though the work appears to be at an early stage and no release date has been set. 9to5Mac said the wording of the listing suggests development has only recently begun.
The listing follows a separate Bloomberg report on Nov. 8 that Apple wants to make Siri, its voice assistant, behave more like an autonomous agent, able to proactively complete tasks without the user starting every step. Mark Gurman, a Bloomberg reporter who covers Apple, wrote that those capabilities could arrive as soon as next year.
Globe TV also reported that Apple’s chief executive, John Ternus, has been pressing Craig Federighi, the company’s head of software engineering, to speed up AI development. Apple has not publicly confirmed that account.
Why it matters
Apple has been widely seen as trailing rivals in generative AI, and the race has shifted this year toward agents that can take over a user’s screen. 9to5Mac wrote that OpenAI, Anthropic and Meta are among the companies now offering their own agents. Meta’s Muse app passed 10 million downloads in about 40 days, according to app analytics firm Sensor Tower, as cited by 9to5Mac and BigGo. Google, meanwhile, began letting its Gemini assistant control Mac apps for subscribers this month.
Those third-party agents run on Apple’s own hardware, which gives the company both an opportunity and a problem. Apple has recently tightened rules on how outside AI agents can reach into its computers:
- Apple said last month it would revise macOS privacy controls around Full Disk Access, a setting that lets software read nearly everything stored on a Mac, according to 9to5Mac.
- The changes appear in the macOS 27.3 beta and give users more granular control over what outside agents can access, the outlet reported.
- Globe TV reported that Apple told developers in October it would require “very explicit user action” before apps could obtain that level of system control.
A home-grown agent would let Apple offer similar features while keeping tighter control over privacy and security, a selling point the company has long emphasized. It also raises questions the listing does not answer, including which devices would support the agent, how much processing would happen on the device versus in the cloud, and what safeguards would govern an AI that can run code and act with a user’s credentials.
What comes next
Apple has not announced a product name, platform or launch window, and none of the reports disclosed how much the company is spending or how many people are working on the project. Apple typically previews major software features at its Worldwide Developers Conference each June, which would be the next likely venue for any public announcement of more agent-like capabilities for Siri or the Mac.

