AppWords Concierge: a new in-app assistant that “preempts search”
The company wants to enable a Google Now-like experience within apps.
Mobile deep-linking platform Deeplink has introduced what it’s calling AppWords Concierge. It operates like a chatbot, designed to engage in-app users, offering a conversational messaging experience that surfaces content, enables transactions and makes suggestions based on context.
The company doesn’t call it a chatbot, however. Rather, they liken it more to Google Now: predictive or preemptive search.
Deeplink describes AppWords Concierge this way: “an in-app assistant, designed to delight end users with preemptive search results and suggestions.” Interestingly, Deeplink argues, “bots are just a new interface for search.” Over time, the company says AppWords Concierge “will learn user preferences, extract intent from actions performed inside the app, and offer help in finding/doing other things.”
The screens below reflect a mock-up of a travel-app experience, where the assistant helps move a user on a path toward a hotel reservation.
Philosophically and practically, Deeplink presents AppWords Concierge to developers as a kind of antidote to mobile search, which it says is broken:
We feel that mobile search is broken, and any means of keyword input is not going to fix it. While companies large and small continue to focus on evolving the user-query search paradigm, we believe the solution lies in a preemptive mobile (search) assistant, that is embedded inside of the apps people use and love, so kinetic intent can be captured by providing users with helpful suggestions.
AppWords Concierge is currently only available in private beta, but the company has nonetheless exposed three tiers of service and accompanying pricing information on its site.
Deeplink said that in its early tests, AppWords Concierge delivered “impressive engagement metrics: ~30 percent increase in session length, ~40 percent increase in page views/session, and a ~15 percent increase in conversions.”
This development is interesting on two levels. Assuming it works and in fact learns over time, it could be a credible in-app alternative search — though not in all contexts and use cases — and address problems of engagement and content discovery. It also illustrates how bots and “conversational UIs” are starting to fan out beyond messaging apps.