The Evolution Away from Conversational User Interfaces
Those digital devices that get you through the day will get less chatty and more focused on working for you.
Conversational user interfaces (UIs) are a growing focus for application developers. Sometimes known as chatbots, digital assistants, or cognitive agents, these UIs have come to mobile devices through Siri, Facebook Messenger, Samsung Bixby, and similar offerings. Powered by natural language processing (NLP) and artificial intelligence (AI), they’ve entered the consumer IoT arena through Amazon Alexa, Google Assistant, and similar voice-activated home-appliance initiatives. And they’re big in the interactive voice response market.
But, in spite of all this activity, conversational UIs, as we know them now, may turn out to be a fad. That thought occurred to me while perusing this recent interview with Rob High, IBM VP and Watson CTO. In it, he parses the practical distinctions, as he sees them, between three terms that are often construed as synonyms: “chatbot” vs. “conversational agent” vs. “digital assistant.” But as he worked through his discussion it was clear to me that the evolution of this “conversational” technology might result in apps that achieve the same ends but without need for much if any conversation.
We can best understand these approaches as a continuum of NLP-equipped microservices with ever greater powers of predictive agency.
For starters, a chatbot is essentially a microservice that is stateless, event-driven, and focused on in-the-moment natural-language query response. The term that High uses to describe this use case is “single-turn exchange,” where each “Alexa, do this” or “Bixby, do that” NLP request-response flow is independent of any others that a user might make now or in the future.
By contrast, a conversational agent engages in a stateful interaction model in which natural-language request-response flows are contextualized within a longer-running engagement. In a conversation, the agent attempts to infer a user’s deeper intention, such as through a sequence of adaptive contextualized voice-prompts, in order to deliver a more satisfactory outcome.
Then there’s the notion of a “digital assistant,” in which the microservice operates in a deeply stateful, engaged, and personalized fashion with each user. High refers to this as a type of “personal butler” that “knows you deeply but is dedicated to just you and serving your needs.” A distinguishing feature of digital assistants is their ability to anticipate what you will need under various future circumstances, make recommendations at the appropriate times, and, potentially, take actions without requiring the user to actively verbalize their consent of each action the bot might take.
All of this suggests that conversational UI technology may be evolving toward what I’ll call the “silent servant” paradigm. Over time, the enabling AI will grow more predictively accurate, users’ trust in these bots will deepen, the technology will become more sensitive to our emotional state, and it will be built into every physical and virtual artifact in our lives. It’s only a matter of time before users will demand that the bots simply go about their duties without unnecessary conversation. To the extent that myriad intelligent assistants tap into a common pool of identity, profile, and other personalization data, they might help users maintain work-life balance across disparate bot-augmented personas.
The evolution of conversational UIs toward silent servants will be driven by customer preference. Users will gravitate toward conversational UIs that spare them from unnecessary chatter. This technology will not scale in our cluttered lives if we’re being constantly interrupted and overloaded by the need to interactively converse with by every gadget in our pockets, homes, cars, and offices.
The most intelligent bots of the future will be those that know how to do their jobs proactively and invisibly without verbal spoonfeeding.
Jim is Wikibon's Lead Analyst for Data Science, Deep Learning, and Application Development. Previously, Jim was IBM's data science evangelist. He managed IBM's thought leadership, social and influencer marketing programs targeted at developers of big data analytics, machine ... View Full Bio
We welcome your comments on this topic on our social media channels, or [contact us directly] with questions about the site.