Amazon's voice activated digital assistant known as Alexa has matured fast since it was first introduced, gaining thousands of new skills and forming strategic partnerships with the likes of Domino's Pizza and Uber. Now Amazon is offering up the same underlying technology that powers Alexa to developers so they can build conversational bots, or chatbots, that understand natural language.
The service is called Amazon Lex. It allows developers to easily build conversational interfaces into any application using voice and text. What developers get with Amazon Lex are the advanced deep learning functionalities of automatic speech recognition (ASR) for converting speech to text, and natural language understanding (NLU) to recognize the intent of the text. By tapping into Amazon Lex, developers can create chatbots and other interfaces that are capable of engaging with humans in a realistic way.
"Speech recognition and natural language understanding are some of the most challenging problems to solve in computer science, requiring sophisticated deep learning algorithms to be trained on massive amounts of data and infrastructure. Amazon Lex democratizes these deep learning technologies by putting the power of Amazon Alexa within reach of all developers," Amazon says. "Harnessing these technologies, Amazon Lex enables you to define entirely new categories of products made possible through conversational interfaces."
This is a fully managed service that Amazon is offering, and one that scales automatically. The appeal for developers is they don't have to worry about the infrastructure. They also don't have to worry about developing a sophisticated artificial intelligence engine, which itself is no easy task.
There is also appeal for Amazon here that goes beyond just dollars and cents. One of the main ways of improving AI is by processing large amounts of data. Through its Amazon Lex service, Amazon can analyze the data that is uploaded to the cloud to further enhance its AI's capabilities, which will both improve Amazon Lex and Alexa.
Amazon's pricing model for this service is based on actual usage. The cost runs $0.004 per speech request and $0.00075 for text requests. If a bot processes 4,000 speech requests and 1,000 text requests, the developer would be charged $16.75. However, developers can try out Amazon Lex for free—for the first year, Amazon will not charge for up to 10,000 text requests and 5,000 speech requests per month.