The voice app killer was subject of some debate (Twitter-bate) this week after an article by Bloomberg outlining that Amazon Alexa has 80 000 skills but This platform is almost four years old, and you can’t point me to one single killer app. Voicebot also published an article related to it with a slightly different take: killer apps are already here, what comes next?
What we are likely to see is a series of killer apps some of which may be localized to a particular surface. For example, voice interactive navigation may be a killer app for the car, but only a minor benefit for smart speakers and inconsistently used on mobile.
As per usual with Voicebot analysis, I believe this is factually true, but allow me to elaborate.
The Bloomberg post inspired several Tweets, to one of which Alexa Chief Evangelist Dave Isbitski replied that Alexa IS the killer app. My response to his tweet triggered an interesting thread by Dave that you should absolutely go and read because is not reflecting the echo chamber of voice tech opinions. I don’t think there’s necessarily a contradiction between the killer app argument and Alexa (or more generally, digital assistants) as the main apps. My point is that in looking at the killer app for Alexa we are losing sight of what’s in front of us, but we might have been too focused on the current state of affairs: the app model. What’s the killer app for your operating system? Is it the clock, the kernel, the task manager? You wouldn’t have to choose among them, because the software that matters in your operating system is the operating system itself. You could, of course, argue that you rather have a kernel than a clock, to the extent of functionality, now to the user that a kernel is a black box the perspective might vary.
The app model
Smartphones and particularly the introduction of the App Store more than 10 years ago by Apple created the app model: a distributed platform for mobile applications with instant reach to users with an internet connection that fundamentally shifted how users interacted with programs. Driven by the iPhone’s touch interface as the primary way to perform tasks, users could do simple and eventually very complex tasks from a device in their pocket, available everywhere.
I was reminded by Ben Thompson how Tanenbaum defined an OS as something that does two things, it abstracts hardware, and it creates an app model. This classic definition can be rephrased for voice technology as the operating system that abstract hardware and the app model. For voice-enabled digital assistants, users don’t need to know that a third party skill it’s an application, as indeed the discovery problem 3rd party skills show. Increasingly, new operating systems will not only be abstracting the hardware but the software as well. Serverless functions provide a clear example to this, but the killer example is digital assistants. I believe the conversations we’ll have with our devices in the years to come will be as Adam Cheyer, founder of Viv Labs has envisioned for Bixby -even if Bixby itself is still lacking in user engagement and available applications – , direct with the assistant which will be the one routing requests, that will abstract the user from knowing that they need to talk to <skill_name> to get the weather for the next week and will simply provide the weather without ever disclosing that it called <skill_name> to reply.
The model is fundamentally shifted from the app model in mobile: Apps that allows to DIY do it yourself to an operating system that is voice-first and DIFY (Do It For You) model. The implications of content ownership are profound, users only know about the assistant and the assistant has knowledge and context of user’s preferences that the third-party apps do not directly have.
Paradigm change is hard
We are living in a paradigm change era similar to the mobile one but with far greater reach and consequences. It’s a paradigm change in how users interact with devices, a new interface, a new OS. When mobile came in 2008, a lot of companies discard it as a fad, a thing that will not affect them, something they could catch up later on. What happened changed the course of technology and shaped some of the biggest tech companies today – and consolidated Apple as one of the most valuable companies in the world. Perhaps the clearest example of a failure in adopting mobile strategy is that of Microsoft, to which “they were fundamentally unequipped to compete in”. The rest is history as they say.
What’s a killer app anyway?
Now to the extent I mentioned there was not necessarily a contradiction between Dave’s point and the need for a killer app for Alexa. The killer apps are those that drive consumers to the platform in a way that wouldn’t exist without them. However, smart assistant installed base and usage has grown significantly YoY because the killer app is the interface and the operating principle: voice and Do It For Me DIFM.