The Danger of a Device-Based Approach to Assistants

With Amazon Echo and Google Home now both in the market, dedicated device-based personal assistants have been in the news quite a bit lately. I’ve been called several times by reporters asking me whether Apple, Microsoft, or other companies need to have similar devices. But if there’s one thing the reviews and my own experience with these devices has taught me, it’s there’s a danger in equating your digital assistant with a device.

Note: this isn’t a review – but Carolina did a great review of her early experience with the Home yesterday.

Equating the assistant with the device

Amazon’s Echo began life as the only home of its personal assistant, Alexa and, although Alexa is now available on several other devices, my guess is the vast majority of users still equate the assistant with the device. Google, meanwhile, has made Google Home the entry point for its own Google Assistant and, for many people, Home is the only place they’ll be able to experience the Assistant for now, given the low uptake of the Allo messaging app and the high barriers to smartphone switching.

The downside here is, as people equate the assistant with the device, they will also equate failures by the assistant with failures of the device. When the entire purpose of a device like Echo or Home is to act as an assistant, to the extent the assistant fails to do its job, the device becomes useless. This is, importantly, very different from the likely reaction to failure by Siri or Cortana, which are mere features on devices that do much more. If we’re unhappy with Siri’s performance, we might well fall back on other ways to interact with our devices or be more selective in the scenarios for which we use Siri rather than the touchscreen because we have options. We may also choose to try again at a later time when the software has been updated because the assistant is still there on the device we’re using for lots of other things. But a device whose sole purpose is to be a good voice assistant and fails at that one job fails entirely and we will likely be tempted to return it or, at the least, put it away.

An assistant trapped in a box

The other challenge with equating an assistant with a device is users can easily have the sense the assistant is effectively trapped in the box. This is very much the case with Alexa, which doesn’t yet exist on smartphones or mainstream wearables. Leave the house and you effectively leave Alexa behind where she can’t do you any good at all. The Google Assistant has been designed with a much broader eventual footprint in mind but, for now, Google has limited its availability to this home device, a smartphone that will sell in small numbers, and an obscure messaging app. That’s a deliberate decision on Google’s part to sell more devices but it also means the Google Assistant will be similarly trapped in the home for many users.

Even when they venture out of the home, having many people’s first experience with the Google Assistant be tied to a larger device with far-field voice recognition technology risks disappointment when people then try it on a smaller device which is less effective at interpreting commands. Conversely, if Apple or Microsoft ever bring their existing virtual assistants to in-home devices like the Echo and Home, users may be pleasantly surprised at the improved voice recognition and will also enjoy a more mobile experience with assistants also present on their other devices.

An assistant that needs an assistant

The other thing that’s struck me again as I’ve been using the Home over the last few days is how important the companion mobile app is, something I noticed with the Echo as well. It’s almost like my assistant needs an assistant, not just for the initial setup but also for other subsequent experiences as well. One of the great advantages Google has in this space is the massive trove of web data it has to tap into but, of course, much of that data is visual in nature – having the Home read you a recipe all in one go is a terrible experience if you actually want to cook something and, of course, image search is also completely useless on the Home. You need the companion app to make sense of those things but, if that’s the case, then why not just use your phone? And, once you’re using your phone for some of these interactions, why not use it for all of them?

I’ve found that what you really want in quite a few of these interactions is to have voice interaction as the primary interface but with some kind of screen as a confirmation or feedback interface as well. A phone (or an Apple Watch or Android Wear device) gives you that combination but the Echo or the Home don’t. One of the frustrations I’ve had with the Home is that, when it fails, it’s often not clear whether that’s because it misinterpreted what I said or because it simply hadn’t been programmed to deal with the request, even when properly understood. A screen of some kind can eliminate that ambiguity.

Best as part of an ecosystem of devices

It’s early days in the history of these assistant-speaker devices for the home and I’m sure we’ll see some meaningful advances in the future. But I’m still finding the utility and performance of these devices is more limited and frustrating than transformative in my home. And I’ve been reinforced in my belief these devices have to be endpoints, not the endgame when it comes to virtual assistants and that such virtual assistants will only be truly effective when they’re part of an ecosystem of devices and not just a single device.

Published by

Jan Dawson

Jan Dawson is Founder and Chief Analyst at Jackdaw Research, a technology research and consulting firm focused on consumer technology. During his sixteen years as a technology analyst, Jan has covered everything from DSL to LTE, and from policy and regulation to smartphones and tablets. As such, he brings a unique perspective to the consumer technology space, pulling together insights on communications and content services, device hardware and software, and online services to provide big-picture market analysis and strategic advice to his clients. Jan has worked with many of the world’s largest operators, device and infrastructure vendors, online service providers and others to shape their strategies and help them understand the market. Prior to founding Jackdaw, Jan worked at Ovum for a number of years, most recently as Chief Telecoms Analyst, responsible for Ovum’s telecoms research agenda globally.

11 thoughts on “The Danger of a Device-Based Approach to Assistants”

  1. “I’ve found that what you really want in quite a few of these interactions is to have voice interaction as the primary interface but with some kind of screen as a confirmation or feedback interface as well.”

    Confirming my view that these assistants are at their best when they can listen to us rather than when they talk to us. Because (as you note) robot voices don’t flow naturally so it can be hard to understand them speak a long thing. Because audio is low bandwidth compared to visual. Because robot servants, like real servants, work best when they are unobtrusive, and a robot voice is the opposite of unobtrusive.

    Being able to command a computer without lifting my hand: awesome in the right contexts. Having a computer that can talk: not nearly as useful as it’s made out to be in the movies.

  2. “You need the companion app to make sense of those things but, if that’s the case, then why not just use your phone? And, once you’re using your phone for some of these interactions, why not use it for all of them?”

    I think Home has an advantage that it is a multi-user device, while Google Assistant on your phone is personal. Also Home potentially could be configured to “always listen” (after Google allows soft-selection of “trigger words” instead of OK,Google). I think Google Home with phone personal assistant can be a good combination since it combines a ubiquity with a mobility.

  3. Hopefully we’ll circle back to some kind of glasses, because I still think they’re the less disruptive, obtrusive and cumbersome way to both get queries/orders into an Assistant and get the answers from it. Regular-looking and camera-less, please ;-p

  4. Apple already has an assistant that is available everywhere thanks to Airpods. It’s very sad that so many tech reporters seemed to have missed the point that Airpods will one day soon allow communications with all your devices from your home, to your car, to walking about with your iPhone, to wherever you work with your iPad or MacBook.

    The key difference between it and environments with multiple platforms is that you don’t have one user friendly seemless experience. Instead of learning multiple devices, operating systems, and using various access points, you can just learn Siri and iOS to have easy access to all.

    1. I doubt people will wear that 24×7. To start with, it doesn’t have the battery for it. then it looks rather silly (I’m not betting on fashion evolving for it, Moto’s pods at least attempted to be discreet). If you got to fish the case out of pocket, then the pod out of its case, then put it in, then do all that in reverse… a phone is quicker and easier.

      1. Airpod is simply the 1.0 version. The airpod gets fine standby time. You can leave one in the charger and keep one in your ear and switch when you have to. Using that model, most users could get through the day with few switches.

        As to the fashion, the first version, like the first iPod and iPhone had first to be functional so you can test it out. Keeping one in your ear is no different than having a Bluetooth headset in terms of fashion. I don’t see why most would not y walk around their home, get in the car with it, go to work, and reverse the cycle. As to social settings, then you can take it off just like you don’t walk around a party or club with your bluetooth headset.

        The key advantage airpod has over the competition is the noise isolation cone. As such it leapfrogs all other interfaces that are succeptible to much of the noise of home, car, and work life.

        For those that master Siri dictations and commands, the productivity boost and convenience is what we have all wanted since watching Star Trek the next generation always available voice computer assistance.

        1. Apple does manage to change fashion and suspend common sense (glass phones etc…) where other brands can’t.
          BT earplugs haven’t managed to become socially acceptable up to now though, and they’ve been around for a long time. Even Moto’s thingies which were much smaller while offering very similar services (minus a few years’ of AI advances).

          We’ll see ? I think something will happen once the services rendered outweigh the social/bother cost. Not sure when either side of the equation will reach breaking point.

  5. On the other hand, personifications, even the non-anthropomorphic ones, work well as a UI paradigm. The assistant being a thing, not just the ether, might make it easier to grok. Plus it kind of meshes with use cases, the living room assistant is not *my* assistant, but the family’s, my personal assistant needs to be in a personal device.

    The main challenge is to find a handy object to put the personal assistant in. Watch, glasses, earplugs, phone… I’m sure the assistant will transfer to other devices from time to time, but I’d bet it will be more successful in one device 90+% of the time.

Leave a Reply

Your email address will not be published. Required fields are marked *