The expanding scope of conversational user interfaces
There are many applications for conversational interfaces in business, for both customer and employee users. Input devices include not only smart speakers but also phones, PCs and vehicles. The conversational interface also works without voice; for someone at a keyboard, typing the command is easier, faster and less disruptive than speaking it.
Conversational interfaces offer benefits far beyond hands-free operation. Users need not remember specific applications and commands, but rather the interface can suggest a possible meaning of what the user has asked or can say it does not understand and users can then rephrase and try again. This reduces the burden on the user to learn how to make a command and thereby expands the number of commands available. For some commands, like making an appointment, speaking or typing a command in natural language is far easier to use than any point-and-click or touch-and-swipe interface.
Additionally, when there is a wide range of possibilities from which to select, the conversational interface is supreme. The user need not remember the exact name of the item. Naming the item as the user remembers it, and then correcting as necessary, is far easier than any of the conventional user interface tools for choosing from many possibilities, including hierarchical menus, scrolling lists and incremental search. Items include names of people as well as names of songs, television shows or podcasts for consumer applications along with products, regions, or variables for business users. A well-implemented conversational user interface can make it very easy for businesspeople to query a complex database, and they can do it using a smartphone. Widespread adoption of conversational interfaces in business will drive up the value and utilization of data, leading to more applications that create potentially valuable data.
As conversational user interfaces evolve, they are becoming more powerful. Currently, the popular consumer voice interfaces process user input on servers, so each conversational interchange requires a round trip between the user’s device and the cloud. As devices become more powerful and conversational software improves, there will be a migration of voice processing from the cloud to the edge. Google (Nasdaq: GOOGL) has announced that the next generation of Google Assistant will process requests on the Android smartphone, starting with the company’s Pixel phone. Google claims this change will improve responsiveness by a factor of 10, and will allow a richer vocabulary of commands, more complex back-and-forth dialogues between the user and the device, and greater adaptation of the interface to the needs of each user. TBR believes this local processing will leverage AI capabilities built into the devices and, in the case of the Google Pixel, will use Google’s Edge TPU, the mobile version of the company’s server-oriented Tensor Processing Unit (TPU). With increased power and better security and privacy, this migration of processing to the edge will drive growth of business applications.