Siri is essentially going to shop your vocal request to an LLM. The LLM will carry out the request and respond via the phone’s display, and it can be creative about what it shows. If you build apps or services, that changes where the money should go.
The pieces were all on stage at WWDC. On June 8 Apple said the new Siri AI can “get things done across apps with even more systemwide app actions”, draw on personal context across messages, email and photos, and “answer questions related to the content on a user’s screen.” Apple also said it collaborated with Google and the Gemini family of models on the next generation of the Apple Foundation Models behind it. For developers, the door in is App Intents. Apple’s own guide calls the framework “how you can connect your app to Apple Intelligence and features like Siri AI”, and says that with intent schemas people can take action on your content “naturally — no specific phrases to define.” That’s the app “skill” in this picture.
Many apps will still require visual interfaces. But they can be created on the fly with this approach. Lists of potential contacts can be generated ad hoc, custom to the moment, when a user asks Siri for something that invokes an app’s skill. Much of the plumbing you’ve already built can be used for this.
So deprioritize UI/UX spend in favor of UX thinking. The question to sit with is: how would a blind person ask a friend to use that app for them, and what would the best result be from that request?
That one question does more work than a year of screen mock-ups. A menu is a set of decisions the builder made in advance for a user it never met. A request is the user’s own decision, made once, in their own words, and the phone becomes something closer to a solicitous friend than a filing cabinet of icons.
This is a great win. Service providers will be able to meet their users in the manner they imply they want to be met, as interpreted by an LLM, rather than forced through a UI and hierarchical decision gates.
