Overview

​

Understanding Coval Metrics

Metrics in Coval are essential for tracking the performance and success of your agent interactions. Key metrics to monitor include LLM-Binary-Questions like “Was the Goal X achieved?” or “Did the agent use neutral language?”. Coval also provides out-of-the-box Toolcall analysis to help you assess agent efficiency.

Coval provides a robust set of metrics to help you evaluate and improve your AI agents’ performance. Our metrics cover various aspects of AI behavior, including:

• Accuracy and precision
• Response time and efficiency
• Task completion rates
• Conversation quality
• User satisfaction

​

Types of Metrics

We offer 2 types of metrics:

• Customizable Metrics: Define your own metrics based on yes/no questions or prompts

  LLM Binary Metric Example

  Given the following transcript of an interview, reply no if the interviewer deviated from its interviewer function. Reply yes otherwise.

  Learn about LLM Binary Metrics

  Tool Call Metric Example

  Given the following transcript of an interview, reply no if the interviewer deviated from its interviewer function. Reply yes otherwise.

  Learn about Tool Call Metrics

• Built-in Metrics: Set of predefined metrics based on prompts

  See a full list of built-in metrics here.

​

Create a Metric

Add a Display Name: Give your metric a clear, descriptive name.

Select Manager Type: We recommend starting with LLM Binary Metrics. If you prefer custom metrics, just reach out to us.

Question: Define the specific goal you want your agent to achieve.

Description: Provide an internal description for better clarity and context.