In order to make its world-class content and services ecosystem available to Amazon Echo and Facebook Messenger users, Edmunds needed a solution capable of automatically responding to both text and speech questions in real-time. To bring the solution to life, the approach consisted of these three steps:
- Design a stream processing flow that takes messages from a source such as Facebook Messenger, Amazon Echo, etc.
- Implement an Apache Spark Streaming prototype to ingest messaging data from selected sources
- Develop an Alexa assistant prototype for implementation with AWS