Recently, AWS announced Amazon Transcribe Call Analytics, a new feature of Amazon Transcribe. AWS customers can extract valuable insights from customer conversations like customer and agent sentiment, and conversation characteristics such as non-talk time, interruptions, loudness, and talk speed with this new feature.
At AWS re: Invent 2017, the company first launched Amazon Transcribe, which provides an automatic speech recognition (ASR) service that allows developers to add speech-to-text capabilities to their applications. After the general availability in 2018, AWS further evolved the service with, for instance, the automatic redaction of Personally Identifiable Information (PII) feature last year. And now another feature is added with Transcribe Call Analytics - a machine learning (ML) powered API, allowing developers to analyze call recordings to get turn-by-turn call transcripts and actionable insights, which they can add to their call applications.
Through the Transcribe Call Analytics console, developers can choose any available call category templates and configure some logic. Next, they pick a model type, a general or custom language model, and specify the language settings. Subsequent steps are providing the input location, for instance an S3 bucket, picking the appropriate AWS Identity and Access Management (IAM) role, and launching the job. Finally, developers can also start a job created through the console by calling the StartCallAnalyticsJob operation of the API or AWS CLI.
Note that the output job can, for instance, be visualized in a word document by leveraging an available Python 3 application on GitHub.
Source: https://github.com/aws-samples/amazon-transcribe-output-word-document
Julien Simon, an artificial intelligence & machine learning evangelist at AWS, explains the value of the new feature in a recent AWS news blog post:
Based on ASR implemented in Transcribe, Transcribe Call Analytics adds natural language processing (NLP) capabilities specifically trained on customer calls, and optimized to provide highly accurate call transcripts and actionable insights. With a simple API call, developers can now easily add call analytics to any application, and extract customer insights from conversations without having to build AI pipelines and train custom ML models.
Currently, Transcribe Call Analytics is available in 21 languages at the following AWS regions: US West (Oregon), US East (N. Virginia), Canada (Central), Europe (London), Europe (Frankfurt), Asia Pacific (Mumbai), Asia Pacific (Singapore), Asia Pacific (Seoul), Asia Pacific (Tokyo), and Asia Pacific (Sydney). Furthermore, the feature adheres to the same pricing model as Transcribe - after a 12-month free trial, AWS customers pay $ 0.0005 (per second) for the first 250,000 minutes per minute analyzed. More details on the pricing of Amazon Transcribe are available on the pricing page.