This article may contain affiliate links. Learn more on our Partner Program page.
An AI agent that operates the computer like a human: It surfs the web, switches between apps, writes code, and decides for itself which tool it needs. This isn't a leak, but rather Apple itself, in a new job posting for a leadership position in Cupertino.
On November 17, Apple posted the position "Senior Engineering Manager, Agentic AI – Proactive". The company is looking for a manager to lead a team of engineers, applied AI researchers, and systems specialists working on agents for general computer use. The listing is one of the most concrete indications yet of how far Apple's plans for autonomous AI now extend.
This aligns with a report from early November stating that Siri AI is to become an AI agent, and that Apple CEO John Ternus is pushing for this. The overall picture of Siri AI shows what the assistant can already do today.
Key Takeaways
- Apple is looking for a Senior Engineering Manager for „Agentic AI – Proactive“ in Cupertino, advertised on November 17th.
- The agents are designed to perform multi-stage tasks across apps, browsers, command line, code and programming interfaces, coordinating specialized sub-agents in the process.
- The team works closely with the departments for Siri, UI frameworks, and Apple's own AI models.
- The models are intended to run on the devices, under strict specifications for response time, memory and energy.
- Apple cites security, data protection, and permissions as a central part of the design.
What the job advertisement describes
The ad begins with an image of what Apple plans to build:
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.
According to the team, they are already working on such agents. These agents are designed to plan and execute multi-stage tasks while coordinating specialized sub-agents. The future leader will be responsible for the technical direction of the agent platform and will translate the capabilities of partner teams and the team's own expertise into high-performing, reliable, and secure agents.
The tasks include the architecture of a platform that navigates the user interface and selects the appropriate tool and level of autonomy for each task. The advertisement doesn't explain exactly what the „Proactive“ suffix in the title means.
On the device, not just in the cloud
The executive will work with the teams for Apple's own base models to bring high-performance and efficient models to the devices, adhering to strict specifications for response time, memory, and power consumption. Apple lists experience in optimizing models for use on the device as a desired qualification.
The agents will be trained and tested in simulated environments: replicated apps, websites, and developer tools where many runs can be performed in parallel. Together with the evaluation and data teams, a definition will be established for when long, continuous tasks are considered successful, measured by actual user results.
Safety as a condition
The more autonomy an agent is granted, the greater the risk. The advertisement addresses this openly:
Agents that can run code, use credentials, and act on a user's behalf need strong permission models, sandboxing, and transparency.
Security, data protection, and protection against attacks should therefore be central to the design. The desired qualifications include relevant experience with sandboxing and authorization models for systems that execute code or act on behalf of users. Experience with accessibility features and automation within the Apple ecosystem is also required.
How the ad fits the Siri course
In early November, Bloomberg reported that Apple plans to develop Siri into an agent capable of performing tasks independently. This agent-like development could reportedly begin as early as next year. The fact that the agent team is said to be working closely with the Siri department supports this direction.
The advertisement doesn't describe a finished product. It mentions neither a name nor a timeline, and it's unclear whether the agents will be part of Siri, a standalone feature, or initially for developers.
Why a job advertisement reveals more than a rumor
I find the advertisement more revealing than many a supply chain report. Apple itself describes what a team is already working on, and the picture is ambitious: not an assistant that answers individual questions, but a system that operates programs and executes code. It's striking how much emphasis Apple places on permissions and sandboxing, and that the models are intended to run on the device. This corresponds to the approach Apple already used to introduce Apple Intelligence: processing as much as possible on the device.
I expect Apple to introduce these capabilities gradually. The expansion of Siri AI is already proceeding similarly: With iOS 27.3, Apple is initially bringing the new capabilities to numerous of its own apps. Developers have been able to connect their apps since the beta, but can only submit them to the App Store starting this spring. All articles about the assistant are collected on the Siri AI topic page.
Would you allow an AI agent to independently use programs and execute code on your Mac, or would that be too much relinquishment of control? Tell us in the comments which task you would first entrust to such an agent.





