As with many technologies (see VR, 5G, etc.), natural language conversations were oversold well before the underlying technology could meet the goal of transforming user experiences. Companies branded relatively simple voice recognition and keyword mapping as "artificial intelligence," and users quickly became frustrated with dead-ends while interacting with chatbots or virtual assistants. Furthermore, on the supply side, companies had not meaningfully organized data to support indirect or "fuzzy" matching (e.g., there may not be a direct answer about GDP growth, but here are some points on inflation and unemployment that may help).

But in this next decade, I expect partnerships between a few major players (Google especially) and many targeted entrants (e.g., Drift, Expressive) to refine the NL formula and radically transform the way we interact with many applications. The most significant impact of conversational platforms will be the further democratization of technologies across a wide user base. Navigating through cluttered menus or disparate articles is not an experience that leads to quick adoption, clear value capture, and prolonged use. Developers have been aiming to alleviate this and make applications as self-service as possible, and the panacea may be "natural" conversation.

This will certainly upend business models that rely on antediluvian interactions (such as click-bait advertising), and unfortunately, there is significant risk that the major AI players will capture much of the value through hosting, data access, and IP licenses. But upstarts that deeply understand their customers' conversational workflows will have an opportunity to leapfrog slow-moving players in long-standing industries (e.g., BI tools) or create completely new solutions that fundamentally rely on this new form of interaction (e.g., Woebot). The one major headwind is the potential loss of jobs, but as with manufacturing, skills will migrate from providing the services to overseeing and improving them.