iOS 27's Public Beta Just Put the New Siri in Millions of Hands. If You Build iOS Apps, Your App Intents Window Is Open Right Now.
iOS 27's Public Beta Just Put the New Siri in Millions of Hands. If You Build iOS Apps, Your App Intents Window Is Open Right Now.
On July 17, Apple released the public beta of iOS 27, which means the redesigned Siri left the developer sandbox and landed on regular people's phones for the first time. The new assistant understands what's on your screen, searches your mail, messages, and photos, answers general-knowledge questions, and takes actions across the system (running on-device models with Private Cloud Compute for anything heavier). Public beta is a lot of installs. Enthusiasts, testers, and a chunk of Apple's most engaged users are now living with this Siri months before the fall general-availability release.
If you build iOS apps solo or in a tiny team, the interesting part isn't the Siri demo. It's that the window to be discoverable by this assistant is open now, during the beta, and how much of your app it can actually reach depends on work you either did already or haven't started: exposing your app's core actions to the system so Siri can find and invoke them.
The assistant is a new discovery surface, and it reads intents
Here's the shift worth internalizing. For years, "Siri support" was a nice-to-have you bolted on if you had time: a couple of shortcuts, a voice trigger, mostly ignored. An assistant that understands on-screen context and performs cross-app actions changes what that integration is for. When Siri can look at what a user is doing and offer to complete a task, the apps it can actually drive are the ones that have described their capabilities to the system in a machine-readable way. The apps that haven't are invisible to it: the user has to open them by hand, the old way.
That's the reframe: the actions you expose to the system are becoming a discovery and distribution surface, not a convenience feature. If a user asks their phone to do something your app is built for, and your app hasn't told the system it can do that thing, the assistant routes around you: to a competitor that did the work, or to Apple's own first-party equivalent. On a platform where discovery has always been the hard part for indie developers, a new place to be found is worth taking seriously, and a new place to be invisible is worth worrying about.
I want to be careful not to overclaim the mechanics here: Apple's exact routing behavior in iOS 27 is still settling during the beta, and how aggressively Siri surfaces third-party actions versus first-party ones is one of the things worth watching over the next few weeks. But the direction is clear enough to act on. An assistant that acts across apps needs a machine-readable description of what each app can do. Providing that description is the price of admission.
Why the beta is the window, not the GA date
The instinct is to wait for the fall release before doing integration work. That instinct is backwards this time, for two reasons.
First, the users on the public beta right now are disproportionately your early adopters and reviewers: the people who try new capabilities, tell their friends, and shape the first wave of "does this app work with the new Siri?" impressions. Being present during the beta means those users find your app through the assistant while the behavior is novel and they're actively poking at it. Show up after GA and you've missed the group most likely to notice and reward the integration.
Second, betas are where you find out what actually works. Assistant routing, the phrasing users try, the actions Siri chooses to surface: you learn all of it by having your app in the system while real people use it, not by reading documentation in September. Every week you're in the beta is a week of feedback you can't buy later, and a week your competitor who shipped early is already using.
What I'd actually do this week
Audit your app against one question: if a user asked their phone to do the main thing my app does, could the system route that request to my app? Then work backwards from the answer.
Start with your app's three or four core actions: the verbs that are the reason it exists. For a task app that's "add a task" and "show me what's due." For a fitness app it's "log a workout." For a notes app it's "capture this." Those are the actions that need to be exposed to the system as invocable capabilities, phrased the way a person would actually ask, not the way your internal API names them. If you've already done App Intents work, revisit it against the new context-aware behavior: intents written for the old shortcuts model may need richer parameters and better natural-language phrasing to be useful to an assistant that reasons about what's on screen.
If you've done none of this, pick the single most important action and expose that one first. You don't need full coverage this week. You need to not be invisible for the thing you're best at, while the early-adopter wave is watching.
The honest take
I could be early on this. Public betas don't always translate to the behavior that ships, Apple has walked back assistant ambitions before, and it's possible the fall GA Siri surfaces third-party actions far more conservatively than the beta suggests, in which case the discovery upside is smaller than I'm painting it. That's a real risk, and if you have exactly one week of runway this quarter, this might not be where it goes.
But the downside of acting is small and the downside of waiting is asymmetric. Exposing your core actions to the system is bounded, well-understood work you'd want done regardless, and it pays off the moment the assistant can use it. Waiting saves you a few days now and costs you the early-adopter window and the beta feedback loop if the discovery surface turns out to matter. On a platform where getting found is the whole game, I'd rather have shipped the integration and found out it was slightly early than skipped it and found out it was the distribution channel I ignored. The window is open during the beta. Use it.
Author
Lukas
@lukcombinator