Building an agent feels like a gamble without the rules
Apple stands at a threshold that every platform eventually faces: the moment when the tools it hosts begin to act, not merely respond. By opening its App Store to AI agents and weaving them into Siri within iOS 27, Apple is quietly redefining what it means to be an operating system — shifting from a stage where apps perform to an environment where software pursues goals on a user's behalf. The outcome will depend less on the technology itself than on the wisdom of the rules Apple chooses to write, and how much trust developers and users are willing to extend before those rules are fully known.
- Apple is preparing to allow AI agents — software that acts autonomously on a user's behalf — into the App Store, a change that could fundamentally alter how people use their iPhones.
- Developers are caught in a holding pattern: without published guidelines on what agents can and cannot do, building for this new platform feels less like opportunity and more like guesswork.
- The silence around permitted actions — whether agents can make purchases, alter settings, or access sensitive data — is generating friction before a single agent has shipped.
- Apple is racing to resolve that uncertainty at WWDC, where it plans to unveil the concrete rules that will govern agentic AI on its platform.
- The competitive pressure is real: Google, Microsoft, Meta, and OpenAI are all building agent infrastructure, and Apple risks irrelevance if it moves too slowly — or user backlash if it moves carelessly.
Apple is preparing to welcome a new class of software into its App Store: AI agents capable of acting on a user's behalf with minimal human direction. The company is building this capability into iOS 27, with Siri positioned as the central gateway through which users will summon and direct these autonomous applications. Rather than requiring users to open apps or tap through menus, agents would handle tasks — scheduling, purchasing, information retrieval — through natural language alone.
For developers, the prospect is both exciting and unsettling. Apple has not yet published the guidelines that will govern what agents are actually allowed to do, leaving builders uncertain about whether their software will survive App Store review. Can an agent make a purchase? Modify device settings? Access sensitive personal data? The absence of answers is creating hesitation before the technology has even launched.
Apple appears to recognize the urgency. The company is preparing to use WWDC — its annual developer conference — as the moment to translate its vision into concrete policy. The stakes are considerable: guidelines that are too restrictive could push developers toward Android or outside the App Store entirely, while rules that are too permissive could expose users to privacy violations and unauthorized actions.
The broader race is impossible to ignore. Every major technology company is investing in agent infrastructure, and Apple's decision to open its platform is an acknowledgment that standing apart is no longer an option. Whether developers respond with enthusiasm or caution will depend almost entirely on what Apple says next — and whether the rules it writes can honor both its tradition of curation and the ambitions of software designed to act without asking permission.
Apple is preparing to open its App Store to a new class of software: AI agents that can act on behalf of users with minimal human direction. The company is building this capability directly into iOS 27, with plans to deepen Siri's role as the central point of contact between users and these autonomous applications. The move represents a significant shift in how Apple thinks about the relationship between its operating system and third-party developers—one that could reshape the entire mobile app ecosystem if it succeeds.
The company's strategy hinges on integration. Rather than treating AI agents as isolated tools, Apple wants them woven into Siri, the voice assistant that has been a fixture of iOS for over a decade. This approach would let users summon agents through natural language commands, allowing the software to handle tasks—scheduling, purchasing, information retrieval, coordination—without requiring the user to open an app, tap buttons, or provide step-by-step instructions. For developers, this represents both an opportunity and a constraint: their agents would gain direct access to millions of iPhone users, but only if they conform to Apple's vision of how autonomous software should behave on the platform.
But enthusiasm among developers appears tempered by uncertainty. Several concerns are holding back adoption, according to reports from people familiar with the matter. Developers worry about what Apple will actually allow these agents to do. The company has not yet published comprehensive guidelines spelling out which actions are permitted, which are restricted, and which are forbidden outright. Without that clarity, building an agent that passes App Store review feels like a gamble. Will Apple allow agents to make purchases on a user's behalf? To modify settings? To access sensitive data? The silence on these questions is creating friction before the technology has even launched.
Apple's timing suggests the company is aware of the need to move quickly. The company is preparing detailed announcements about agentic AI policies and App Store guidelines for WWDC, its annual developer conference. That event will be the moment when Apple attempts to translate its vision into concrete rules—the dos and don'ts that will govern what developers can build. The stakes are high. If Apple gets the balance wrong, either by being too restrictive or too permissive, it could either stifle innovation or create security and privacy problems that damage user trust.
The broader context matters here. AI agents represent the next frontier in how people interact with software. Rather than apps designed around specific tasks, agents are designed to understand goals and figure out how to achieve them. They can coordinate across multiple services, learn from user behavior, and make decisions in real time. Every major tech company—Google, Microsoft, Meta, OpenAI—is racing to build agent infrastructure. Apple's move to welcome them into its ecosystem is an acknowledgment that the company cannot afford to sit on the sidelines. At the same time, Apple's historical emphasis on privacy and security means the company faces pressure to ensure that agents don't become vectors for data theft, unauthorized transactions, or manipulation.
What happens at WWDC will likely determine whether developers embrace this new capability or approach it with caution. Clear, permissive guidelines could unlock a wave of innovation. Overly restrictive rules could push developers toward Android or toward building agents outside the App Store entirely. Apple has built its platform on the premise that it can curate the user experience while still allowing developers to build ambitious things. The question now is whether that balance can hold when the software in question is designed to act without asking permission.
The Hearth Conversation Another angle on the story
Why does Apple need to open the App Store to AI agents at all? Couldn't they just improve Siri on their own?
Because agents are fundamentally different from traditional apps. An agent needs to coordinate across multiple services, make decisions, and act on your behalf. That's not something a single app can do well. Apple needs developers to build that ecosystem.
So what's stopping developers from building agents right now?
Uncertainty about the rules. Developers don't know what Apple will allow agents to do. Can they spend money? Access your contacts? Modify settings? Without answers, building an agent feels like building on quicksand.
Why would Apple be vague about this? Wouldn't clarity help everyone?
Probably because Apple is still figuring it out themselves. Agents are new territory. The company has to balance openness with safety and privacy. That's genuinely hard to get right.
What happens if Apple's rules are too strict?
Developers go elsewhere. They build on Android, or they build agents outside the App Store entirely. Apple loses control of the ecosystem it's trying to create.
And if the rules are too loose?
Then you get agents that steal data, make unauthorized purchases, or manipulate users. Apple's brand is built on trust. One major breach could undo years of work.
So WWDC is the moment of truth?
Exactly. That's when Apple tells developers what's actually possible. Everything before that is just speculation and anxiety.