Speech recognition and a bunch of APIs

2018-12-07 1 min read

    Google’s Duplex is clearly impressive but what I actually want to see is the frontend of Duplex hooked up to the various booking APIs. For example, instead of actually making a call to a restaurant on my behalf it would just use the various booking sites already out there, such as OpenTable, to make the reservation. From the caller’s experience the two are indistinguishable since you’re just using your voice but on the restaurant side there’s no need to talk to someone to do something that can be done through a series of API calls. The calling approach is incredibly impressive and makes sense when a business doesn’t offer any online booking functionality but many do and involving a person introduces overhead.

    Imagine combining the speech recognition power of an Echo or a Google Home with the investment that’s already been made in the various APIs. Google is the leader in voice recognition and should be able to find a way to map speech to intent that can be executed without actually needing to interact with a person.