Credits: This lab is heavily based on the following two AWS blog posts and the respective code:
In this lab we will establish a processing pipeline that will automatically process any new call recording. This will employ a few additional AWS services and features we haven’t used before today. And it will obviously also include some code that takes care of what we did manually so far.
The processing pipeline that is created in this lab assumes stereo wav files as they are produced by Amazon Connect. In case you use Amazon Polly for producing your audio files, there are two things that need to be changed after the CloudFormation stack is complete. These are explained in Using Amazon Polly.
We will first have another look at the architecture, which we will trigger for automatic creation as next step.
Then we will manually trigger the pipeline with uploading a call recording. After that we will integrate the pipeline with our Amazon Connect instance so that every new call recording will be replicated into the pipeline for processing automatically.
At the end, we will also have an introduction into analytics and visualizations of our results.