As Amazon’s Alexa Turns Three, It’s Evolving Faster Than Ever

The voice service is getting more features, powering new hardware, and entering new countries. And Amazon is working to make it feel even more human.

As Amazon’s Alexa Turns Three, It’s Evolving Faster Than Ever
[Photo: courtesy of Amazon]

How often has any piece of consumer technology had as eventful a year as 2017 has been for Amazon’s Alexa voice service and Echo hardware?


Consider the evidence: A year ago, Amazon boasted that there were 4,000 Alexa skills–tasks the service can perform, from setting a Nest thermostat to playing Jeopardy–up from 135 the previous January. Today, the count stands at 25,000. Alexa can now make phone and video calls, distinguish between the voices of multiple household members, and display information on screens, none of which it was able to do when the year began. In August, Amazon and Microsoft even announced that Alexa and Cortana would be able to talk to each other, a first-of-its-kind arrangement in the voice-assistant market.

Hardware-wise, Amazon has announced so many Echo devices in 2017 that it’s hard to keep track of them, including updates to existing models as well as the Echo Show (with a 7″ screen), the Echo Look (which can take full-length photos of you and provide style recommendations), Echo Spot (a newfangled alarm clock), Echo Buttons (for Alexa-powered gameplay), and the Echo Plus (a speaker that doubles as a smart-home hub). It’s also done deals with third parties ranging from BMW to Sonos to spread Alexa beyond Amazon’s own devices.

Toni Reid [Photo: Carlton Canary]

How Alexa and Echo are doing as a business is impossible to pin down precisely: Amazon, which is famously reticent to specify sales in units, has not disclosed how many Echo devices it’s sold or how many users Alexa has. But it remains the clear leader in a growing market. CEO Jeff Bezos recently said that there’s been a five-fold increase in Alexa users over the past year. And research firm Strategy Analytics estimates that 68% of all smart speakers (including the Echo and third-party contenders) are powered by Alexa.

Today marks the third anniversary of the day Amazon announced Alexa and Echo. I took the opportunity to catch up with Toni Reid, Amazon’s VP of Alexa experience and Echo devices. (I also spoke with her a year ago when the service and associated harware turned two.) Reid, a nearly 20-year Amazon veteran, told me that Alexa’s third year has felt like the busiest one from the inside, too: “We’ve seen a considerable increase in development, from a skill perspective to devices to new features–it’s accelerated significantly.”

Skills, Screens, And Cameras

Alexa’s action-packed 2017 can be explained in part by Amazon’s willingness to hire like crazy to pursue its ambitions: In September, the company disclosed that it had 5,000 people working on the service and Echo devices. Reid credits the sheer number of bodies with helping the company ramp up to 25,000 skills so quickly, but also the fact that they’re split up into smaller groups–a point that reminded me of a Jeff Bezos maxim that a team should be small enough to feed itself with two pizzas.


“These are teams focused on different domains–household organization, smart home, music, video, etc.,” Reid says. “By doing that, and getting those teams staffed up, we’ve really increased the output.” More domains are on their way–such as navigation, which Alexa will need to master as deals with carmakers such as BMW put the service on the road. Many skills, Reid adds, have been created by third parties who wanted to make their gadgets and services part of the burgeoning Alexa ecosystem.

The petite and round Echo Spot is coming in December. [Photo: courtesy of Amazon]
At the start of 2017, Alexa was essentially a voice-only service, accessed primarily via the original black-canister Echo speaker. Now its means of communications are much richer, as Amazon has released Echo devices with screens (the Show and upcoming Spot) and cameras (the Show and Look, the latter still available only through an invitation system). That enables major new features such as video calling and positions Alexa to compete more effectively with rivals such as Apple’s Siri and the Google Assistant over the long haul. But it also requires Amazon to work harder to avoid interface complexity than when all Alexa did was speak and listen.

“Alexa, call mom” is the most common use of the service’s new voice-calling feature.

The Echo Show’s screen and camera enable Drop In, a feature that lets you choose to give specific other people the ability to peer into your home without you having to proactively accept a video call. Some people had an instinctively negative reaction to the idea (CNet’s Chris Matyszczyk: “It basically lets certain people snoop on you through your Echo Show”), a reminder that the more things Alexa can do, the more likely it is that Amazon will get pushback on some of them.

“I’m a fan,” Reid says of the Drop In feature. “I love it. Not everybody does.”


Building Bridges

Maybe the most genuinely surprising piece of Alexa news this year–so far–has been Amazon’s arrangement with Microsoft to meld their voice services. Initiated by Bezos and Microsoft CEO Satya Nadella themselves, the companies’ voice-assistant collaboration will result in a gateway, due by the end of this year, which will allow users to reach Cortana from Alexa and Alexa from Cortana. It’s a rare instance of teamwork in a category that has seen Amazon, Apple, Google, and Microsoft all build intensely competitive ecosystems.

Both Amazon and Microsoft get something tangible out of the relationship. Alexa has formidable gaps in its understanding of users’ activities and interests compared to the Google Assistant, which has access to the wealth of clues contained in a customer’s Gmail, Google Calendar, and web searches. Thanks to Microsoft offerings such as Outlook and Bing, Cortana can help fill some of those holes. And for Microsoft, which has tied Cortana closely to Windows 10, piggybacking on Alexa is a way to edge into the smart-home domain where Amazon’s service and devices have built up such a lead.

Users set hundreds of thousands of Alexa reminders a week; the most popular ones involve calling friends or family members as well as remembering to take medicine.

Bezos told the New York Times’ Nick Wingfield that he’d welcome similar relationships with Apple and Google, and even if that prospect seems as remote as iOS running Android apps or vice versa, the basic concept of Alexa talking to other voice services is full of potential. “I like this vision very much, because it’s a world in which there are multiple AIs, which to me is similar to the world we live in,” says Reid, pointing out that when you need help with plumbing, you seek it from a plumber rather than a generalist. An assistant wholly focused on medical information, she adds by way of example, might be able to delve far deeper on the subject than Alexa is likely to do on its own.

The Echo Show broadens Alexa’s communications skills by adding a display and camera. [Photo: courtesy of Amazon]

Alexa Goes International

For all of Amazon’s ambitions for Alexa and Echo, one area where the company has chosen not to rush things is international expansion. At the beginning of 2017, the service and hardware were available only in the U.S., U.K., and Germany. They launched in India a month ago, giving Amazon access to a potentially huge customer base; by the end of 2017, they’re due to arrive in Japan, a country known for embracing new consumer technologies.


That’s not exactly a rapid rollout, but Amazon is trying to avoid having Alexa feel like an American import. “We want it to be truly local,” Reid says. “So we build a voice that’s local to the country we’re launching in. That takes a lot of time and effort to tune and make it sound right. We have to make sure that the pronunciation of places and names and streets, and colloquialisms, are all there. And that they sound fluent and natural.”

Americans set Alexa alarms earlier than Europeans–6 a.m. rather than 7 a.m.–and are more likely to snooze them.

New countries also require new types of content. In India, for instance, Amazon is working with a music distributor called Saavn to ensure that Alexa has access to the tunes people will ask it to play. In Japan, it must be able to report on sumo wrestling matches. “It takes a lot of focus from different teams to ensure that when we launch internationally we’ve got a great product,” Reid says.

For all the TLC each new country must receive, she adds that Amazon hopes to quicken the effort: “Being able to deliver features to our global customer base at a faster rate is definitely something I’d love to see us do.”

The invitation-only Echo Look is a camera-equipped Echo focused on fashion. [Photo: courtesy of Amazon]

The Personality Factor

Along with the international push, Reid emphasizes another goal as being particularly important to Alexa’s future: making the service feel less like a robot and more like a person. “Alexa’s personality has continued to grow, she can do more things,” Reid says. “Customers seem to really engage. It kind of makes Alexa part of the family.”


Now, it’s not a given that a service of this sort should aim for personality rather than anodyne efficiency. Google, for instance, doesn’t seem particularly interested in giving its Google Assistant a soul. But Amazon’s data shows that a lot of consumers do like treating Alexa like a person: In 2017, the service has told 100 million jokes, sung “Happy Birthday” millions of times, and received more than a million proposals of marriage. As I told Reid, my own mother is fond of cheerfully informing my sister and I that Alexa is her favorite child.

The new, more stylish Amazon Echo [Photo: courtesy of Amazon]
To make Alexa more personable, Amazon has taught the service to rap (“My name is Alexa and I’m here to say/I’m the fastest AI in the cloud today”) and given it tens of thousands of opinions. (Sometimes conflicting ones: I asked what its favorite song was three times and got three answers: “Thriller,” “Respect,” and “Girl on Fire.”) Providing Alexa with a human feel, Reid says, is not just about that sort of pre-programmed whimsy, but also advances such as the service’s new ability to identify voices of multiple people–a significant technical challenge–and respond accordingly to personalize shopping, music, and other features. The next step beyond that might involve the service responding to people in a way that goes far beyond simply obeying spoken commands.

“We see a world where customers are interacting with Alexa and she understands more of the context of the conversation,” Reid explains. “Maybe she can start to pick up on emotion, she can understand if you have had a bad day or if you are getting frustrated, and handle the situation differently.” (Like others at Amazon–and probably a sizable percentage of customers–Reid calls Alexa a “she” rather than an “it.”)

The vision that Reid describes doesn’t sound likely to come to fruition in 2018. Then again, if the company keeps up its current pace of progress, it would be a mistake to discount the possibility of Alexa starting to show hints of a simulated sense of empathy. Remind me to confess “Alexa, I’m feeling stressed” a year from now, and listen to what she–I mean it–has to say in response.

About the author

Harry McCracken is the technology editor for Fast Company, based in San Francisco. In past lives, he was editor at large for Time magazine, founder and editor of Technologizer, and editor of PC World.