A new window is opening in connectivity. In the early stages of electronic telecommunications history, the industry evolved around connecting people – human to human. Later, computers were connected with other computers, and the growth of Internet connectivity followed. In recent years, the focus of attention has been on connecting other things to the Internet as well (IoT). Now, entering yet another phase, machines are learning to understand human speech, enabling humans to connect with things, at a distance.
“The future market size is easily in the billions of dollars.”
Technology giants are spearheading the development. Both Amazon and Google have produced speakers for the home that understand voice commands. This development is triggered by the improved quality of voice technologies – which in turn is driven by the introduction of deep learning in the branch of AI called Natural Language Processing.
In effect, the speakers function like “cell tower” equivalents that can connect other things with people as well. In the home, people will not only communicate with the speakers themselves, but through the speakers also with other objects. This is a platform strategy, connecting many customers with many products. Product manufacturers can either decide to develop their own speakers and AI (less likely), or partner with a service provider like Amazon or Google (more likely).
The opportunity lies within the high number of things that will be connected. In a typical household one may find a series of products that are commonplace today that employ buttons. Amazon is already integrating its speakers with many such products, including: light dimmers, thermostats, TVs, Wi-Fi routers, dishwashers, cooking ovens, garden sprinklers, cars, carports, and more. Arguably, a smart home makes a lot more sense with a voice interface.
The future market size is easily in the billions of dollars. The market size for providers can be estimated with a back-of-the-envelope calculation: Total number of households x Average number of connected objects in a household x Average revenue charged per connected object. In the US alone, there are over 130 million households. If every household on average contains 8 things that can be connected, the US total exceeds 1000 million devices. Annual profit per object depends on the pricing model – which can be agreeable, given that each device is used several times a day, most days of a year. Note that other business models may be employed as well – Google may come to prefer an ad-centric model, for example.
Additional billions of revenue are waiting for Amazon and Google. RBC Capital Markets recently estimated that Amazon’s Alexa can increase the company’s revenues with $10 billion by 2020. Half of these billions come from sales of the Echo speakers themselves. The additional $5 billion come from increased e-commerce revenue per customer, as customers increasingly shop from Amazon through voice driven sales.
What can telcos do to capture a piece of this growing pie? Entering the growing market requires significant capabilities. AI is a source of sustainable competitive advantage, being both rare and costly to imitate due to the knowledge intensive process of producing it, and the limited pool of experts with the required know-how. Telcos that aren’t already moving in this area are already at a disadvantage. However, some telcos have responded to the challenge, including Deutsche Telekom and Orange; the two incumbents have joined forces to create an AI assistant branded “Djingo”.
Personally, I am intrigued by the new opportunities. As humans, our primary “API” for communicating with other humans is language. Opening this connection towards things as well opens a realm of new opportunities. Imagine a world without interfaces, where things can intuitively understand us. Not to mention, what are the opportunities (and challenges) when things begin to talk to each other, using language? In the end, moving from buttons to a voice interface may prove to be just as influential as our transition from Morse code to the telephone, back in the 1870s.
