๐น Step 1: Selecting an Agent
The first step is to choose the agent you want to evaluate.Navigate to the Agent Evaluation section in Lyzr Studio and select the desired agent from the list.

๐น Step 2: Generating Test Cases
Once an agent is selected, you can generate test cases to validate its performance.You will be prompted to Name your Test Case Group. This helps in organizing and categorizing test cases for future reference.

FAQ_Bot_Evaluation
or OrderTrackingAgent_Tests
.
After providing a name, click on Generate Test Cases. The system will automatically create relevant test cases tailored to the agentโs purpose.
๐น Step 3: Running the Test Cases
After generating test cases, you can run them against the selected agent.The results of these runs will be displayed, showing the agentโs responses to each test case along with a summary of performance.

๐น Step 4: Generating Improvements
One of the key benefits of Agent Evaluation is the ability to generate improvements.Using LLM-powered analysis, Lyzr suggests ways to enhance the agentโs accuracy, coverage, and overall effectiveness based on the test case results.

โ Benefits of Agent Evaluation
- Systematic Testing โ Create structured test groups to validate your agentโs functionality.
- Organized Workflow โ Manage multiple test case groups for different scenarios.
- Actionable Insights โ Automatically receive recommendations for improvement.
- Continuous Enhancement โ Keep refining your agents with every evaluation cycle.