Apple opens Siri AI to developers for multi-step app workflows

Expanding Siri's capabilities

Apple is making its AI-powered version of Siri more accessible to third-party software makers. Starting with the iOS 27.3 beta, developers can connect their own applications to Siri AI to perform more complex actions. This move allows for deeper integration than previous versions of the assistant offered.

The new system enables developers to provide both content and specific actions directly through App Intents. By using these building blocks, apps can be woven into the broader AI ecosystem. This allows Siri to act as a bridge between different parts of a user's digital life.

As BornCity reported, Apple is also expanding these capabilities across its own platform. Fifteen system apps, including Mail, Safari, and Notes, are being updated to support these new features. This ensures that core Apple services benefit from the same multi-step logic as third-party apps.

How the new 'Skills' work

The technical foundation for these new capabilities relies on a concept called 'Skills.' These are defined in Markdown files and allow the AI to link several individual steps into a single workflow. Instead of performing one isolated action, Siri can now sequence multiple tasks to reach a final goal.

Developers can use a specific Swift macro, @AppIntent(.tool), to declare existing app functions as tools for the AI. This tells the system which actions are available for it to use. To keep data safe, Apple has included specific authentication policies to protect sensitive information during these processes.

Siri AI performs these tasks by looking at the user's personal context, processing natural language, and using screen recognition. Because of this, users do not always have to name a specific app in their commands. If an app is already open or a live activity is running, Siri can understand the intent based on what is happening on the screen.

Real-world use cases

To illustrate how this works, Apple describes a scenario where a user wants to send a favorite pair of sneakers to a partner. The AI would first identify the specific product from the user's saved items. It would then look up the correct recipient in the device's address book.

Once the information is gathered, the AI would automatically send the data through a messaging app. This entire process happens as a single, fluid interaction rather than a series of manual steps. It demonstrates how the AI can navigate different app functions to complete a complex request.

Other examples of these multi-step workflows include managing email distribution lists or cleaning up browser history. Users might be able to ask the assistant to delete files from a specific time period or recall a sent email. These actions require the AI to understand the sequence of events needed to fulfill the user's request.

Availability and limitations

While these tools are now available in the developer preview, they are not yet ready for the general public. The preview is currently restricted to development builds within Xcode and internal TestFlight environments. External TestFlight participants and regular App Store users cannot access these features yet.

Apple expects the final release of iOS 27.3 to arrive next month. However, regular App Store submissions using these new interfaces will not be allowed until the spring. It will take several more months after that before the first consumer versions of these apps hit the market.

There are also several hardware and regional restrictions on the new service. It requires devices that support Apple Intelligence and is blocked for users under the age of 13. Additionally, the service is currently unavailable in the European Union for iOS, iPadOS, and watchOS.

The business of AI

Apple's move into deeper AI integration comes as the company views these services as a primary growth driver. CEO John Ternus noted that Siri handles over 2.5 billion requests every day. The services sector recently reached a record high of $31.6 billion in revenue during the last quarter.

However, the transition to AI-heavy services comes with significant costs. Apple has noted that rising memory prices and the high computing costs of Siri AI are impacting gross margins. The company expects margins to fall into a range of 45 to 46 percent in the upcoming December quarter.

To keep up with the demand for these features, Apple is investing heavily in research and development. Recent spending on R&D climbed by 32 percent to reach $45.8 billion. These investments are intended to support the growing infrastructure needed for the next generation of AI-driven user experiences.