Summary

In this chapter, we learned how to build an NLP solution to accelerate customer service efficiencies using Amazon Comprehend's Topic Modeling feature, Detect Sentiment feature, and by training our own custom classifier to predict routing options using Comprehend Custom Classification before hosting it using Comprehend real-time endpoints. We also saw how we can leverage the flexibility of powerful and accurate NLP models without the need for any ML skills. For your enterprise needs, Amazon Comprehend scales seamlessly to process millions of documents, provides usage-based pricing, supports batch inference, and with autoscaling support for real-time endpoints, you can manage your inference request volumes and control your inference costs effectively.

For our solution, we started by introducing the customer service use case, the inherent challenges with the way things are set up currently, and the need to perform automated routing and sentiment detection to control the...