iOS 27 escaped the developer world today with the launch of its first public beta. I’ve been testing the new operating system since early June, looking for issues and seeing if it lives up to the hype Apple promised in its keynote.
This year’s iOS upgrade is called the Snow Leopard update. That is, it focuses less on new features and more on fixing what’s broken and speeding up processes throughout the OS. App launches, photo search results, and AirDrop transfers should all be faster. The Messages app now supports inline replies and end-to-end encryption for RCS messages. Liquid glass is more refined and has improved readability around hard edges and text. These are all great updates, especially for those with older iPhones. But the biggest and most anticipated change in this update so far is that Apple has finally actually shipped the improved Siri AI in the form of an opt-in beta program. And I think Apple actually did it this time. At the very least, this ultimately laid the groundwork for a successful version of Siri.
Last week I was wondering if I would have time to go to a free concert in the city. The show was 4 hours long and had 3 acts, but I only really wanted to see 1 act. I couldn’t find the band performance order on the event page. So this was the perfect opportunity to pressure test Siri AI. So I swiped down from the top of the screen and asked, “In what order are the bands playing?” Siri spun the new little wheel for a few seconds and told me exactly the information I was expecting. So the band I wanted to see was the last one playing.
This made it impossible to open the browser for most things
The promise of the new Siri AI is to change the way you use your phone. In the past, you would go into an app and tell it what you wanted to do (call a car, set a timer, order lunch, etc.). Here, you first say what you want to do, and Siri AI will try to take care of the rest by looking at all available apps and information. So when I asked about a concert, Siri looked at the content of the web page, searched the web for the answer, and presented it to me. There was no need to jump through browser tabs or look at the band’s Instagram page. It was just there.
I’ve been using Siri AI for a month now and it has surprised me in many ways. On the first day I tested the beta at Apple’s development conference, I was able to ask, “Can you add the WWDC briefing to my calendar?” Siri looked at my email, parsed the data, and added six separate events with the correct times to my calendar. Note, it can only be added to mine apple Calendar, but I’ll explain why in a moment.
These interactions changed my brain chemistry a bit. Now, I almost always use Siri first to see if it can perform the action I need or answer a simple question. It’s easier, faster, and more fun to swipe down from the top of the screen and type in the prompts, so I’ve virtually eliminated opening the browser for most purposes.
I think the on-screen recognition was probably the most helpful for me. You can ask Siri about what’s on your screen, so you don’t have to tap everywhere. Even better if you can take action from on-screen recognition, like adding an event to your calendar or directing you to an on-screen address.
Siri does what I ask more often than I expected, and it feels like magic when performing somewhat complex tasks. But when you hit a wall, you’re reminded why it takes a little work to get to a world where Siri “just works.” do not have Working is the exception, not the norm.
Siri’s ability to turn requests into specific actions appears to be one of the main areas where Apple is still working out issues. For example, if you’re looking at a concert page and ask, “Remind me to buy these tickets when they go on sale,” we’ll create a reminder that simply says, “Purchase these tickets when they go on sale.” I had to be careful about saying “buy a ticket for…”. this I used the “when they go on sale” trigger to have Siri actually look at my screen and search the web for ticket availability. And most of the time, when I asked Siri to “route” me to a location, nothing happened, but it was able to “direct” me to a location. I imagine word correlation might improve over time as more data comes in, but it can be frustrating if the whole point is to be able to talk to the computer in natural language without relying on uttering keywords.
Currently, if you’re using the Siri AI preview in iOS 27 beta, you can only access new Siri features in the Apple app. If you live within the Apple ecosystem, everything is fine. Your data probably lives in messages, emails, photos, and when you want to take action, you add to-do items to reminders and notes. I use several of these apps every day, and when they work together, like adding a list of email events to a calendar, it feels like a glimpse into the future. But when you ask Siri, “When did Daniel say you were free to play?” Dota? ” In that case, Siri doesn’t know anything. Daniel and I only communicate through Telegram, and the system doesn’t have access to Telegram.
There are two main updates that developers need to implement for Siri AI support: Entities and Intents. Entities represent the types of data that you can include in your app, such as photos, recipes, playlists, and notes. When an app adds an entity to Siri AI, Siri knows exactly what kind of data it can retrieve from that app for use in personal context. Conversely, intents tell Siri what you can actually do with that data, such as play, save, or delete. Entities and intents in your app allow Siri to control and retrieve data from them, and Siri’s semantic layer handles understanding the user’s words. Entities and intents in your app allow you to control and retrieve data with these semantics.
And if most of your digital life is outside of Apple apps, it will take time for the apps you want to work properly with Siri AI. It will never happen during public beta. Developers can now build entities and intents against the iOS 27 SDK, but as a user you’re out of luck until the SDK itself is out of beta. In fact, you won’t be able to push Siri AI updates to your apps until the full version of iOS 27 is available in the fall.
Apple relies on many Developers will have to update their apps to support the new Siri, but it’s not like this hasn’t happened before. Every time Apple added important new features like dark mode or iPad apps, developers had to work to make them work properly. I think the difference is that most of the updates so far directly impact the user experience within the developer’s own app. Now, Apple is asking developers to update their apps to improve the Siri experience.
The developers I spoke to said that while they’re very excited about these new features, updating their apps to support them is quite an undertaking. “The conceptual challenge for developers is to create comprehensive support for every screen and feature within an app,” says Matthew Cassinelli, who worked at Workflow, which was later acquired by Apple and became Shortcuts. “However, the move to an agent-based model allows dedicated apps to dynamically display relevant data, making them more useful for users who don’t frequently open a particular application.”
An example of this is the LookBack: Contact History app. This app will show you who you have recently added to your phone’s contacts. It’ll also show up if you forget to open the app and ask Siri, “Who did you meet at a conference last week?” You can use the app to view information.
For small developers who aren’t building the five or six apps that most people use every day, the argument that Siri can help surface small apps that users frequently forget to open is understandable. But the bigger question in my mind is whether a company like Google would be interested in making Siri even more capable. Google makes most of its revenue from showing ads, and if Siri could simply display that information at the top of the screen instead of forcing you to open Gmail to find the information you’re looking for, Google would lose out on that revenue. That being said, Google is doing essentially the same thing with AI Overview, so it’s clear that they’re preparing for a world where AdSense revenue isn’t their primary source of revenue.
I believe the primary motivation for Google to adopt full support for Siri AI is consumer choice. If one email app has the best support for Siri AI and Gmail doesn’t, I’m more likely to use the other app. I want Siri to continue running through these experiences to derive my personal context and take actions. If you start to run into roadblocks in terms of what an app can do, you might start looking for another app. Whether Google’s own base is strong enough to lock in most people is another question.
This all remains a work in progress. Not only is this a beta OS, but it’s also a beta Siri. And that’s just part of the full experience we’re promised in the future. That’s impressive, but what really matters is still: Developers must add support for Siri AI to realize its full potential. And Siri AI needs to correctly route and investigate requests far more often than it stumbles. So far I’ve been surprised at how much it can do.
By the way, the concert was amazing.
Photo by David Immel/The Verge
