Apple wants to do the work between the apps
Apple spent years making it easy to find an app for almost anything. Its next AI project could make switching between them somebody else’s job.
A now-removed engineering job listing described an agent capable of browsing, controlling apps and writing code. The role involved coordinating work across Siri, Apple’s foundation models and the software frameworks behind its interfaces. anitab.org Job Board
The ambition is broader than answering a question. An agent could take a goal, work through several steps and use different tools along the way.
Development remains early, with a launch eyed as soon as late 2027. Meta and Anthropic shares weakened in late trading after the report, reflecting the prospect of another formidable competitor. uk.finance.yahoo.com
Apple has not announced a finished product. Still, the direction fits the challenge facing John Ternus, who became CEO on September 1: keeping Apple central to the way customers use their devices as AI changes those habits. Apple
The valuable position is where a request becomes an action
Think about arranging a trip. The useful part is getting the itinerary, reservation and calendar entry to line up without doing the same work three times.
If an assistant handles that coordination reliably, it could become the first place a customer turns. App makers would then compete partly over whether that assistant can find and use their services.
Apple already supplies some of the plumbing. Its App Intents framework lets developers expose actions and content to Siri, including support for moving content between apps. developer.apple.com
That gives Apple a practical advantage: it can develop the assistant alongside the system through which applications make their capabilities available.
It also sets the permissions for outside agents. Bloomberg’s report describes 14 steps to enable ChatGPT’s computer-use tools in a macOS beta, up from four. uk.finance.yahoo.com
Permissions serve a legitimate purpose when software can access files, use credentials or act on someone’s behalf. But extra setup can also discourage adoption. Security design and competitive access are becoming closely intertwined.
For rival AI companies, a highly capable model is therefore only part of the contest. Getting customers to grant access—and keep using the service—could be equally consequential.
Follow the AI Race From the Data Center to Your Devices
Apple’s smaller construction bill tells only part of the spending story
Apple has taken a different financial route from the companies building enormous AI campuses.
Its fiscal 2026 cash spending on property and equipment fell to $9.04 billion, from $12.72 billion. Meanwhile, research and development rose to $45.77 billion from $34.55 billion.
That additional research spending exceeds $11 billion. Apple does not separate out the AI portion, but the figures make clear that a smaller capital budget does not mean spending stood still. images.apple.com
Partnerships matter, too. Apple and Google agreed in January to a multiyear collaboration using Gemini models and cloud technology for the next generation of Apple Foundation Models. The companies said Apple Intelligence would continue running on Apple devices and Private Cloud Compute. blog.google
Apple can therefore compete for the customer relationship while relying on outside technology. That arrangement also leaves room for a model provider to benefit from Apple’s success.
An agent introduces a harder cost question. Multistep work can require repeated model calls, checks and retries. Running more of it on devices could limit cloud costs, but the balance will depend on capability, performance and usage.
Buying computing services can reduce the need to build infrastructure directly. It still leaves a bill.
Apple’s opportunity is to make its devices more useful—and harder to leave—without letting the cost of serving those users outrun the benefit. Its job listing specifically sought expertise in efficient models constrained by memory, power and response time. anitab.org Job Board
That is where the product story meets the financial one. A convincing demonstration would show useful work completed reliably, with clear permissions and an approach to computing costs that can scale beyond a carefully staged example.
Disclaimer
Wall Street Access is an independent financial publisher. Wall Street Access is not a financial advisor. No company mentioned has compensated Wall Street Access for the preparation, publication or distribution of this article. This content is provided for informational and educational purposes only and is not investment advice or a recommendation to buy, sell or hold any security. Investing involves risk, including the possible loss of principal.
Prices, market capitalizations, financial figures and other market data are based on information available as of the date and time of publication and may change without notice. Information is drawn from public sources believed to be reliable but is not guaranteed to be complete or accurate. Opinions and expectations expressed are our own and may change as new information becomes available.
Readers should conduct their own due diligence and contact a licensed financial professional before making any investment decision. Past performance does not guarantee future results. Any material ownership interest or conflict of interest, where applicable, will be disclosed separately.