MLA-C01 Practice Question 44: Call Center Audio Analytics
AWS Certified Machine Learning Engineer – Associate (MLA-C01) · Test 5 · question 4 of 10
A telecom company stores recordings of its customer support calls in Amazon S3 as stereo audio files, with the agent on one channel and the customer on the other. Managers want a transcript of each call, customer and agent sentiment across the call, and a list of issues the customer raised. The company has no labeled data and no ML team available, and it wants the results within a month.
Which approach meets these requirements with the least effort?
Show answer and explanation
Answer:
Amazon Transcribe Call Analytics is a managed AI service built for contact center audio. A post-call job on a recording stored in S3 returns a transcript separated by agent and customer. It also returns sentiment analysis for both participants throughout the call, call characteristics such as talk time and interruptions, and detected issues, outcomes, and action items. Its sentiment analysis works out of the box with no training data, which fits a company with no labels and no ML team (option B).
- A is wrong because Amazon Comprehend analyzes *text*. It cannot accept an audio file, so the calls would need transcribing first, and generic key phrases are not the same thing as the issues a customer raised on a call. - C is wrong because Amazon Lex is for building conversational bots that act on a user's intent in real time. It is not an analytics tool for recorded two-party calls, and replaying recordings through a bot captures neither sentiment nor who said what. - D is wrong because it needs labeled training data, ML expertise, and weeks of work, none of which the company has, to rebuild what a managed service already provides.
Reference: Post-call analytics - Amazon Transcribe