AI on Your Wrist: Why Smart Watches Need AI to Succeed
Bzzzzt. It’s the doorbell. Today’s the day. 2 weeks ago you eagerly queued up on Apple.com to place an order for the Apple Watch Series 2, and you’re ecstatic. “This is the one for me!” you say — Apple has finally figured out the bugs and added the features you wanted.
Two weeks later, the only metric your Apple Watch tracks how often you open and close the random electronics drawer that became its new home.
Sound familiar? When the Apple Watch first debuted in 2014, CEO Tim Cook described it as 3 distinct items:
- An innovative and intimate way to communicate directly from your wrist.
- A comprehensive health and fitness device.
- A watch (duh).
So far, the Apple Watch has basically become a fitness tracker and luxury watch, but has failed to “integrate itself seamlessly into our lives”. So what’s missing then? The answer: Artificial Intelligence (AI).
“Apple Watch already has AI with Siri!”
Siri hasn’t become a significant part of our daily lives mainly because Siri is still, in essence, a “command taker”. You tell Siri to do something, and it does it (sometimes). You ask Siri for some information, it provides it (if you worded the request just right). Siri does not have the ability to see the “big picture” of all the apps and features in the Apple Watch to create value for the wearer.
Siri lacks the back-and-forth conversation that is essential to creating a seamless AI/Human interface.
The perfect wearable experience.
So, what does the perfect wearable experience look like, then? There’s a few marked differentiators between the Siri of today and tomorrow. Here’s what that will look like:
- Getting the big picture: Siri needs to be able to look at the data provided from multiple different sources (which in a way it already does), but more importantly, find patterns in that data and make predictions on your behaviour on the fly. This mimics human behaviour; when we need to accomplish a task, we look at a wide array of data inputs and make assumptions on the next step to take. For example, an executive assistant may recognize that a CEO has an appointment for 9AM, but receives a flight update that he is arriving 30 minutes later than expected, the EA can handle the situation accordingly (reschedule the appointment). This is really hard for any computer to do right now.
- Conversational UX: We will be able to interact with Siri in whatever way suits us best. One of my favourite apps, Quartz, has got this right. You receive your news from Quartz in a conversational “chat” format. Quartz provides a summary of news stories of the day, focusing on the points that are most important to you. If you are engaged in the story, you send back messages to Quartz to tell you more. This is arguably the biggest opportunity for Apple Watch intelligence. Imagine being able to interact with different apps and updates through one-touch responses, pre-programmed with the actions you’re most likely to take.
And really, that’s it — those two simple (yet difficult) changes will launch the real wearable revolution. Here’s a few more examples of what this could look like:
A day in the life.
In the morning: You wake up perfectly refreshed, with your Apple Watch strapped to your wrist. You notice you wake up a little later than you were supposed to, but that’s okay — Siri saw that you had a rough sleep and gave you a bit more time to sleep in before waking you up and dimming on the lights. Siri saw the only tasks you had to finish up in the morning were not time-sensitive so it rescheduled them to later in the day. Crawling out of bed, Siri sends you an automated message — “Traffic is particularly awful today. You’re probably better off taking the train, shall I find you the fastest route to the transit? Or do you want to take your chances with Uber?” You tap “Uber” and Siri sends a slack message to your workgroup channel that notifies your team your expected arrival time.
During the workday: Walking into work, Siri recognizes your location and wifi network, and immediately switches from the fitness watch face to the utility watch face, which halts all unnecessary notifications and provides background info on your upcoming tasks for the day. You receive an email in your inbox, and before reading it, Siri says “John is asking for a meeting on Monday to go over design mock-ups. You’re unavailable at the time he’s requesting, should I see if he’s available 2 hours earlier?” You tap “yes”, Siri tries to reschedule with John’s calendar, and you move on with your day.
After work: Time to hit the gym, head home and have dinner. Siri again switches watch faces automatically, queuing up your favourite workout routine and asking if there’s anything in particular you want to improve. Throughout your gym workout, Siri compares your workout times and intensities to other Apple Watch users in your base. Siri sees you’re displaying behaviours that show that it’s likely you’re going to give up soon. Siri sends you a message that says “Workout getting a little tough? The gym you’re working out in just offered you a free personal training session, I’ll find a time during your normal workout routine to schedule you in. Keep it up!” Siri sees you leave for home, sends a command to warm up your oven, and queues up a few recipes of items you might have at home. “Expecting company?” Siri asks. You tap no, finish dinner, and hit the hay.
Wearables will continue to be glorified fitness trackers until AI becomes conversational.
On their own, Wearables are simply smartphone mirrors, tracking minor fitness data and sending notifications and interruptions to more of your body. However, with better analysis of the huge amount of data already available, an AI can become very intimate with your routines, behaviours, likes and dislikes, and ultimately free you to do the work and play you want to, without getting bogged down in the details. The intersection between user interface and AI will continue to blur, until finally — all interactions we have with technology will be conversational.
Feedback is always welcomed, and if you can think of other examples, please leave in the comments below.