Talk a little, type a lot - Will conversational interfaces survive Siri and Alexa?

Enterprise conversational systems

I wonder if you are happy with the voice-driven app that your IT group developed to provide you with a digital assistant or process-savvy robot? I bet no one can answer that question in the affirmative. But if Deloitte is right, we can be sure that we will start seeing more and more attempts at enterprise conversational systems in the next several years.

How will we fix these lackluster voice systems that we are stuck with today? Whether we consider consumer-facing conversational interfaces or enterprise-transforming ones, where is the technological improvement going to come from that will move the machine from a pretty good radio tuner or weather reporter to an excellent colleague in a digital economy?

The future of innovation

In past eras of computing, inventive startups have been the source of innovation that moves the industry forward. Nuance Communications, for example, has made a $2 billion-a-year business out of rolling up various small voice-centered firms. But in our current era of digital transformation, with the rise of analytics and new expectations for knowledge-driven systems, voice recognition, natural language generation, and all the component technologies required to produce believable and helpful conversational interfaces, the bar has been raised. The ability to understand and participate in a conversation strains all the AI technologies underpinning our machines today.

For the next generation of conversational computing, it is hard to avoid the conclusion that the only companies that have enough researchers, enough processing resources, enough motivation, and, above all, enough data to deliver the much- needed improvements are the consumer giants we see today: Amazon, Apple, Google, and possibly Microsoft and Facebook. To their credit, they have seen the future in voice for many years already, and they have released products that help show their developers what is bad about them.

All we are looking for from the labs of these giants, I would argue, is to enable the third huge contribution of conversational interfaces—encouraging the machine to disappear. And that is really what we want. There are lessons from history that teach us that as technology matures, it disappears. Nobody bought early automobiles, for example, because they wanted to climb outside and spin a crank to get the thing to start! We wanted to hop in the vehicle and get to our destination with as little mechanical distraction as possible. Conversational interfaces have the potential to bring the key fob to computing.