V-Spark Online Help

Examining Agent and Client Interaction

When V‑Spark processes audio, it is divided into agent and client speakers. V‑Spark helps you explore and audit agent information for things like politeness and emotion, the presence of specific phrases, compliance with requirements for content in each call, and so on. For mono audio, V‑Spark uses a process called side classification to label speech as coming from either an agent or a client. For stereo audio, the channel that contains the agent speaker is configured at the folder level.

The next few sections illustrate the basic mechanisms that V‑Spark provides to examine agent performance and how to work with that information. These sections use an application called Agent Scorecard that was designed to increase the efficiency of the agent evaluation process by automatically detecting the presence or absence of phrases that correspond to QA form requirements. The results of this application can be used to get an overall view of agent performance, pinpoint performance categories that require improved training, and compare scores for individual agents.


Refer to the V‑Spark Review and Analysis Guide for information about the different views available for exploring agent data, application templates such as the Agent Scorecard, and using all of the options on the File Details page.