Android is Using AI to Fulfill the Promise of Windows Phone (Premium)

Anyone still pining away for Windows phone needs to pay attention to how Google is evolving Android with AI capabilities. And no, this platform isn’t somehow “catching up” to Windows phone. It’s succeeding where Microsoft failed.

I still hear from people who miss the innovations that Microsoft brought to its now-dead mobile platform. And I still rarely hear from those who actually suffer along with a Windows phone, as if by denying themselves they are somehow making a point.

I get it. I was the original Windows phone enthusiast, and I wrote the very first book about this platform. And while I never insulted your intelligence by pretending everything was fine, as certain shameless enthusiast blogs did, I held on longer than I should have. It was just hard to believe that ideas that good—and backed by the combined might of Microsoft and Nokia—could fail.

Belatedly, however, I realized that those ideas weren’t actually all that good. And that the many reasons Windows phone failed can all be traced back to those ideas not resonating with the companies that would—or would not—produce the apps and services that ran on Windows phone.

These are the same ideas that Windows phone enthusiasts still cite when mourning the loss of the platform. Live tiles and hubs, key among them. Both spoke to a desire on Microsoft’s part to respond to the iPhone, not by copying it as Google did with Android at the time. But by rethinking how people might use smartphones and trying to make a better user experience.

I wrote “trying” there. Back in 2010, I was convinced that these ideas were, in fact, “better.” What I didn’t understand at the time—what Microsoft didn’t understand—was that this design was hostile to the companies and brands that it needed to make Windows phone a success. What it didn’t do was design a smartphone platform that was better for the entire ecosystems. That’s why it failed.

Hubs are the more obvious problem, so let’s start there.

Hubs were special apps that provided integrated experiences that tied together functionality from multiple online services and standalone apps. The Photos hub, for example, let you browse your photos on your phone, photos from any number of connected online services, and photos that your friends and other contacts were sharing from social media services, all in a single place.

This sounds like a fantastic and user-centric idea. After all, technology is supposed to make life easier. If you want to view, share, or comment on photos, you just think … Photos! What could be easier? Contrast this with the normal experience, even today, where you must think about the specific app first, remember where it’s located on what home screen, and then launch it. Oops, wrong app. Maybe it’s this other one. On and on it goes. I call this “whack-a-mole.” And you do all the work.

The problem with hubs is that the services that need to hook into this system in order for it to work at all—Flickr, Google Photos, Facebook, Instagram, and others, in the case of the Photos hub—all refused to participate. Facebook, for example, wants you to think about its brand, and not about photos generally. It wants you to engage directly with it, using its app, so that it can monetize that usage, usually through ads. Hubs may sound good in theory. But in reality, they are anti-brand. It’s no wonder all of these companies ignored Windows phone. (Facebook was on board only briefly, and my understanding is that it was compensated by Microsoft.)

Live tiles and their “at-a-glance” functionality have the same basic problem: Facebook’s standalone app for Windows phone—to stick with the same example—can show you little snippets of updates from friends and family from outside the app, which can be useful. But you’re still outside the app, and thus outside of Facebook’s ability to monetize you. This functionality, while useful for users, is hostile to brands.

Put simply, Microsoft’s good ideas were only good ideas for people. But they weren’t fully-realized to respect the brands that those people, and Microsoft, needed on the platform. So they were, in effect, not very good ideas. What Microsoft should have done was figure out a way to improve on the “whack-a-mole” user experience while providing its app and services partners with a way to promote their brands.

That is exactly what Google is doing today with Android. And, yes, it is using a quickly-growing set of AI capabilities to make it happen.

Let me provide two examples. These are new features in Android 9 Pie that surface app functionality from outside of the app while preserving the underlying brands. They’re opt-in, so the developers of those apps have to specifically support these features. And unlike on Windows phone, they will do so.

The first is App Actions.

“With App Actions, your app can be recommended to users as a way to fulfill their needs – at the moment they need it the most,” the Android Developer website explains to app makers.

Without getting too far into the developer weeds, app makers can register their apps to respond to a growing collection of intents that appear throughout the system. For example, I wrote recently about another new Android 9 Pie feature called Overview Selection, and how different options can appear when you select text representing a location (Google Maps), a phone number (the Phone and Messages apps), and so on. Those are intents for which a specific app is responding with an action.

These things appear throughout Android, in Google Search, the Google Play Store, the Assistant, and on the Home and All Apps screens. Here are two examples, from Google:

You don’t always have to manually trigger an App Action. On the new All Apps screen, for example, you will find a row of your most-frequently-used apps right at the top, followed by a row of two App Actions tiles.

These change according to usage: In my case, I’ve recently had a group text and was listening to Audible, so perhaps I’d like to continue one of those. Both provide what we used to call a “deep link” on Windows phone. That is, they don’t just launch the app. They go to a specific place in the app. In the case of Messages, if I select that tile I will go directly to the group text.

The second related AI-based feature in Android 9 Pie is called Slices. This one isn’t technically available yet, but Google says it will arrive in the OS in the coming weeks.

Slices are like a special kind of App Action that provide user interfaces—complete with whatever branding the app maker requires—from an app from outside the app. With Slices, apps can overlay a part of their user experience and functionality above whatever other app the user is currently using. Or, at least they will: At first, Slices will only display over Search.

The example Google uses all the time is for the Lyft car service: You search for Lyft and when you select the Lyft app icon in the search results, a Lyft slice appears on top of the Search app. (Or, it will: When you do this today, the Lyft app just opens normally.)

The Lyft slice, in this case, is both interactive and useful. It provides links to destinations like Home and Work, and provides pricing and duration estimates. If you select one of these destinations, the ride is ordered.

Accessing app functionality by searching for the app’s name isn’t exactly smart, to my mind. But that’s not really the point: For now, Google wants an easy way for developers to figure out how they might surface app slices over other apps. And they can configure this functionality for a time in the future when it will be available in more places.

Before moving on, I’d like to address one related issue that seems to be a blocker for some in the Microsoft community: Their desire for whatever mobile platform they use to be just like Windows phone. Or as much like it as possible.

This is misguided. Features like live tiles and hubs made a certain amount of sense when they were part of the platform and were broadly supported by apps and services. And while one might replace the stock launcher on their Android handset with a version that emulates Windows phone to some degree, I recommend looking beyond the past and exploring more modern user interfaces.

And that’s what Microsoft is doing, by the way. You’ll note that the Microsoft Launcher for Android does not provide live tiles or hubs. Instead, it represents a more recent rethinking, on Microsoft’s part, of what a modern mobile device interface can be, and how it can best help the user. Part of that vision includes a PC integration strategy that may or may not pay off. (We’re still waiting on Timeline integration, for example, so it’s hard to know how useful this will be.)

Point being, change is hard, but in this case, it is worth paying attention to what Google is doing with Android and in making an effort to change your behavior. Windows phone is never coming back, but I’ve already moved to a place where I don’t want or need it to. Android today is already plenty functional, and it provides first-class apps and media ecosystems, things that were decidedly lacking on Windows phone. And yes, it is also almost instantly malleable, so that you can change how the system looks and works if you do find the stock interfaces to be lacking. But do yourself a favor and check out the new stuff first. You might be surprised by how good it is.

I’ll be writing feature focus articles about App Actions and, when available, Slices in the future.