Not Everything that has a Voice is an Assistant
If what we have seen in the first couple of days of CES is a taste of what is in store for us in 2018, we can expect to be talking a lot. Most of the time, probably more like shouting commands right, left and center, to devices scattered around our house, in the office and the car.
What last year was an “Alexa takes all” show turned into a stage for all vendors to show off either their support for Alexa and Google Assistant or to announce their own assistant. But as more vendors jump on the bandwagon, it is essential that a distinction is made between a voice interface and an assistant. While they might seem to be the same, they are not and making it clear what is which will secure its success.
An Assistant is an Investment
For an assistant to be helpful you will need to invest time into it. First, at least for now, the user will need to learn how to talk to it. As much as voice is a very intuitive user interface, current voice assistants just do not communicate like a fellow human being. Lack of context limits most exchanges to a set of simple questions and answer. Like a real assistant, a digital one needs to get to know you, which means to know your preferences and information such as calendar appointments, access to apps you use regularly and the devices you might want it to control for you. It’s a true learning curve that will require time even for the smartest assistant.
Of course simple tasks that are more like commands like turn this on, play that, remind me of this, or search queries require little knowledge of us on the assistant part. However, when you want the assistant to be proactive and start doing things for you without you asking, like a true assistant would do, that is when knowledge Is power. To some extent the smarter the assistant the less I should actually need to talk to it.
The Ubiquitous Nature of an Assistant
In the home, my assistant should be pretty ubiquitous to be useful. This is why speakers have been such a focus for Amazon, Google, Microsoft and Apple. Being able to control my devices and ask questions from wherever I am in the home is key to build engagement and ultimately dependence.
At CES and even leading up to it, we have heard of other devices such as TVs, refrigerators, light switches and even showers all coming soon with an assistant inside. Soon our homes will be seeded with many devices able to assist us and that could potentially have a voice.
As we are at the very beginning of this journey, I can understand why vendors are trying to cover all corners as it is yet unclear what device will be the Trojan horse into someone’s home. So for instance, someone might not be buying a smart speaker but be happy to have an assistant integrated in the new TV they are purchasing. I would argue, however, that there is a limit for how many devices should have an embedded assistant versus being able to be controlled through an assistant. The difference in my mind is proportional to the value that an embedded assistant will deliver, which must go beyond being able to execute commands. I might want to be able to control my dishwasher with voice but I do not have to engage in conversation with it, the same can be said about my washing machine. My fridge however, could give my assistant access to a lot of information on how fresh my food is to the best temperature to maintain it fresh, to recipes that would use up what I have left in it. In order for this information to be conveyed the fridge should have access to the internet, have a camera and have a voice and of course be smart. Embedding an assistant vs. connecting the fridge to an external assistant seems to be a much more effective implementation.
If adoption of voice enabled devices goes the way vendors are hoping for, we will also have to have a way to manage all these devices that will be voice enabled. This could go two ways. I would either be able to call my assistant something that is device specific or only the assistant in the device I want will respond based on context. So I would either call my assistant for the fridge “chef” or if I ask “what can I cook tonite?” only the assistant in the fridge will answer me. Right now neither of these scenarios is an option. So if I own a TV, a speaker, a phone, and a fridge all assistant enabled, the likely scenario is that at my question “what can I cook tonite?” I will have my TV show me a cooking program, my speaker say “sorry I am not sure how to help with that yet” my phone will say “here is something I found on the internet” and my fridge will actually give me a recipe. Not very helpful!
Voice UI has Value in Itself
There are more devices that will benefit from a voice UI than they would an assistant. The value that a voice first UI will deliver to users could be huge even if there was no full fledge assistant in the device. This is why I strongly feel that vendors should stay clear from using the term Assistant. Roku has recently announced their Roku Entertainment Assistant and immediately the press has been asking wether it will be better than Alexa or Google Assistant. The reality is that such a comparison is unfair because it is not an assistant. It is a voice first UI that will let users ask to play content with their voice. If you have a Comcast remote, you can do that today. I can press a button and ask “Play Scandal” and the TV will show me all the ways I can watch it. This is not an assistant, it is a voice UI that saves me a bunch of steps for which I am very grateful. Should these voice first UIs even have a name? I would say no. As a user all I need to know is that I can use my voice.
If you think I am over complicating this point, look at how hard was for Samsung to pitch Bixby. Bixby started out as a voice first interface but it was called an assistant and because of that the reviews were fairly negative. This mostly was due to the fact that, as an assistant, Bixby did not have access to a deep pool or data and as soon as users started to use it in the same way they would use Alexa or Google Assistant its value was limited.
Differentiating between a voice first UI and an assistant also brings a series of benefits for what needs to be integrated in the device which could be helpful from a price structure perspective.
An assistant should be much more than a user interface and I think this is where the market is struggling at the moment, because assistants are not actually that smart yet. I truly believe the smarter my assistant will be the less I will talk to it because the power of AI will have my assistant do her job which is making my life easier by anticipating my needs.