Analyse call recordings with Watson

5 min read

Analyse call recordings with Watson

Typically the large volumes of calls handled by a call centre are recorded. Calls are recorded for training purposes, quality control or for regulatory/compliance reasons. While organizations often store large volumes of call recordings, they’ve no means to extract value from these calls without turning to humans to actively listen to the recordings. The Call Analytics with Watson asset aims to address this challenge providing the capability to analyse call recordings with Watson.

What is Call Analytics with Watson?

Built as a web application that can be hosted on IBM Cloud, the Call Analytics asset leverages Watson services to gather insights from call recordings. Using Cloud Object Storage, Speech to TextTone Analyzerand Natural Language Understanding, the asset transcribes the call recording, identifies the tone of each speaker through the course of the call and highlights keywords and entities referenced on the call. Storing these insights in a relational data store (Db2 Warehouse on Cloud) allows an organisation’s BI team use familiar reporting tools to extract business value from this otherwise dark data.

sequence of service calls

How do I get started?

The source to build and deploy the Call Analytics with Watson asset is available on GitHub. Simply provision the required set of services on IBM Cloud, follows the build/deploy instructions, then run the Call Analytics web application to derive real-time insights from call recordings.

Be the first to hear about news, product updates, and innovation from IBM Cloud