Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.cekura.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
Metric Types

Scope of this page. This page covers the shape of the value each metric type returns (Boolean, Rating, Numeric, Enum). It does not cover how those values translate into evaluation success — i.e., how to configure whether a metric affects call/run pass/fail, threshold rules, or scoring rubrics. For that, see Rubric.
These are metrics defined by the user. These can be tailored to specific needs and requirements, allowing users to create custom evaluations that align with their unique goals and objectives. We also support generation of custom metrics based on the user’s voice agent purpose. Below is a list of various types of metrics users can pick to evaluate calls.
This metric directly affects the success/failure of a call.
It evaluates adherence to specific workflow steps, resulting in a simple true/false outcome, indicating whether the workflow was followed correctly.
Example - if asked for discount by customer, the agent should offer 30% or less.
This metric does not affect the success/failure of a call.
This metric offers evaluation on a continuous scale, providing a score on a scale of 0-100%
This metric provides quantitative data for straightforward analysis, including metrics like latency, pitch.
This metric categorizes interactions into predefined categories, helping to identify patterns and trends by categorizing data.
Example - Was the customer happy, sad or frustrated?