AT&T has long been working on a voice recognition engine of its own. The two-decades long research has culminated in the shape of its Watson speech and language engine. The engine quite effectively converts spoken words into text. Now, the carrier has unveiled of plans of making the APIs of Watson available to the developers by June this year.
This can prove to be a very useful feature for app developers. Such developers who have been aspiring to build such apps which make use of voice recognition would no longer have to create a speech recognition software of their own. They will simply be able to make of use Watson APIs to accomplish the task.
The first APIs from the carrier will focus on certain areas which include local business search, web search, SMS, voice mail to text, Q&A, U-verse video programming guide and a general-purpose dictation API. AT&T is of the view that voice recognition software has proved far more effective when it is focused on certain categories. This way, the software knows what kind of words it can expect to hear from the user.
Initially, AT&T has plans of releasing the APIs which will be focusing on the aforementioned areas. However, the carrier may release further APIs related to other categories eventually. The company also plans to launch a SDK which will be termed the Speech Kit. It will enable developers to create such software which can discern speech and then send them over a network to convert them into text. The ‘network’ part of the whole equation is unclear yet but we think that the software may somehow use AT&T’s service to accomplish this.
AT&T has already released a Translator App for both iOS and Android to demonstrate the capabilities of Watson but the apps contains a few bugs which need to be taken care of. To be sure, once the developers have Watson’s APIs, they will be able to put them to a lot of uses.