When Apple opened up Siri to third-party apps a couple years ago, the consensus among developers was that the system was still too limited.
Instead of letting app makers build whatever voice skills they wanted, like what Amazon had been doing with Alexa, Apple restricted Siri to six “domains,” including messages, audio and video calls, ride booking, payments, and photo searches. Support for lists, notes, and QR codes arrived last year, but you still can’t use Siri to play music in Spotify, launch a video in Netflix, or order a pizza from Domino’s.
Yet even within the categories that Siri does support, uptake from developers has been spotty. You can send messages by voice with Viber or Telegram, but not Facebook Messenger or Signal. You can take notes in Evernote, but not in Microsoft OneNote, Google Keep, or Simplenote. You can add to-do list items in Todoist or Any.do, but not in gTasks or Trello. You can send payments using Venmo, but not PayPal. When trying to use Siri with third-party apps, “I wish I could” is a common refrain.
Why does Siri support still feel limited? It’s probably due to a combination of factors:
Theory 1: Users aren’t asking for Siri support
The most simple explanation is that app developers have limited resources, and don’t see the point in supporting Siri when users aren’t demanding it.
That’s the rationale offered by Peter Oehler, the chief operating officer of project management app Zenkit. Being able to add a task or check list items with Siri might be useful in theory, but Oehler imagines that these features would only benefit a tiny fraction of Zenkit’s total user base. And so far, even those users aren’t asking for Siri support.
“We’ve gotten feedback from more than 10,000 users of Zenkit, and are exchanging emails and chats with them. Nobody has demanded Siri integration,” Oehler says.
Ryan Hanna, the founder and vice president of product for workout app Sworkit, also isn’t seeing any demand for Siri support from users.
“We focus more on the challenges of our actual users and what they’re going through. For them, getting a workout in, there are no barriers in their way that being able to launch it by voice would actually solve,” Hanna says. “We’ve literally never had anyone tell us, ‘Hey, I wish you were doing this.'”
Theory 2: Apple isn’t pushing for Siri support
Even if the apps you use do support Siri, you’d be forgiven for not knowing it. While Apple’s App Store has feature sections for apps that support other recent iOS features, such as AR, iMessage stickers, and iPad drag-and-drop, no such spotlight exists for Siri apps. And if you ask Siri what apps it supports, all it does is search the App Store for the word “support.” Currently, the only way to find out if an app supports Siri is through trial and error.
At the same time, Zenkit’s Peter Oehler feels that he’d be going out on a limb to support Siri with no guarantee of reciprocation from Apple.
“From the perspective of a software development company like mine, Apple is not—how would I say it—they are not pushing you,” Oehler says.
By comparison, Oehler points to a recent collaboration with Canonical, the company behind the popular Linux distribution Ubuntu. Zenkit worked with Canonical to build a “Snap” that can easily install and update the app across many flavors of Linux. In exchange, Canonical promoted Zenkit on its blog and in social media, and also surprised Zenkit with a banner ad on the Ubuntu website. Even though Linux users make up a fraction of Zenkit’s potential user base, Oehler felt the investment was worthwhile.
“If I’m working with Apple, I deliver something to them, and they will not do anything for me,” Oehler says. “You are absolutely depending on the App Store, or on things you cannot influence, and you cannot cooperate with them. You are just delivering, and maybe you will get big, or maybe you will fail.”
Meanwhile, other companies are doing more to get developers on board with their voice assistant skills. Amazon, for instance, participates in Alexa workshops and hackathons year-round, and also encourages enthusiasts to hold their own meetups. Google hosts its own events and online talks for Assistant developers, and allows new developers to get detailed, custom feedback on their voice Actions. Although Apple provides lots of online documentation for SiriKit, it doesn’t appear to hold any Siri-related events outside its annual developers’ conference. (Apple has also reportedly seen several senior engineers in charge of third-party Siri integration either leave or move off the project, according to The Information.)
Theory 3: Adding Siri support isn’t always simple
When app developers do get around to supporting Siri, they might find that doing so takes a lot of work.
Matteo Rattotti, one of the developers behind the popular writing and notetaking app Bear, says Siri support required some changes to how the app works.
To work with SiriKit, an app must support iOS’s Intents extension, which defines all the ways that the app can interact with Apple’s voice assistant. While extensions are powerful, they can only read an app’s data through a “shared container,” which is separate from how the app stores data by default. Moving data into a shared container can be tricky and error-prone, Rattotti says, and once inside, it can’t be backed up by iCloud or iTunes.
Also, Siri works differently depending on whether it’s being accessed from iOS, MacOS, an Apple Watch, or a HomePod, and developers must make sure to communicate those differences to users, Rattotti says. All of this means adding Siri support can be a complicated endeavor.
“For us, it was just a low-priority task because we’re a very small company, and we had a really big backlog of features to add, and SiriKit requires some architectural changes,” Rattotti says.
Theory 4: Rival tech giants may want to hamstring Siri
One of the most conspicuous holes in Siri’s app support comes from Google. You can’t use Siri to start a Duo video call, send a message in Hangouts, transfer money with Google Pay Send, take a note in Google Keep, or search images with Google Photos.
Other tech giants also have limited Siri support. You can place calls and send messages with WhatsApp, but you can’t use Siri for messages, calls, or payments in Facebook Messenger. You can use Microsoft’s Skype for calls and messages, but can’t take notes in OneNote or search images in OneDrive. You also can’t send messages with Amazon’s Alexa app or search images in Prime Photos.
All of these companies have their own voice assistant ambitions. Amazon has Alexa, Google has Google Assistant, Microsoft has Cortana, and Facebook–at least until recently–was planning to launch a smart speaker with a focus on Messenger features. This is just speculation on my part, but these companies might not be eager to make Siri more powerful by connecting their most popular apps.
Theory 5: Voice control is still young
It’s worth noting that in the cases of Zenkit and Sworkit, they’re not currently supporting other voice assistants either.
Zenkit’s Peter Oehler says he’s considered working with Alexa’s to-do list voice commands–compared to Google and Apple, he sees Amazon as less likely to try and its own competing project management software that competes with Zenkit–but it’s “not on the upper part of our roadmap.”
Likewise, Sworkit evaluated Google’s voice assistant platform last year, but ultimately decided to focus on features that customers were asking for. While he says that voice assistants could be useful–especially for controlling workouts within an app–they’re currently too limited in what they can do.
“If the experience in your app can be just as good without actually looking at the device, that’s when I see SiriKit being better off,” he says.
That’s unlikely to happen anytime soon. Voice assistants still act mainly as extensions for existing apps and services, performing a limited number of functions quickly instead of duplicating everything a proper app can do. And with Siri in particular, those functions are still limited to a narrow set of app categories.
Perhaps as SiriKit expands to more kinds of apps, more people will become comfortable using Siri for those simple commands, and developers will feel more pressure to add their own support. But for now, they’re not in any rush.
“If you’re not Facebook or the really big ones, you can wait for how the market will develop,” Zenkit’s Peter Oehler says. “We are not losing anything if we wait.”
This post has been updated to clarify the terms of Zenkit’s collaboration with Canonical.