After conducting a test run, a detailed analysis of the test results is crucial to assess the performance of the AI agent and identify potential areas for optimization. Test results show whether the AI agent is reacting as expected or if adjustments are needed.

​
Viewing Test Results

  • In the Test runs overview, select the desired test run and click on the Arrow to the right.
  • The detailed view of the test run is displayed, including status, tested AI agent, runtime, and test result.
  • The overview also shows whether the parameter transfer was successful and if there were deviations or errors.

​
Analyzing Test Results

  • By clicking on the Arrow to the right of the individual test results, the Sidebar with the test details opens.
  • There, the results are divided into two categories:
    • Received: The actual response of the AI agent.
    • Expected: The defined correct answer by which the received result is compared.
  • If there is a deviation, it can be analyzed why the AI agent has responded differently than expected.

​
Accepting or Further Analyzing Test Results

  • If the received answer is correct, it can be accepted as a new test result.
  • If there are deviations, it should be checked whether an adjustment of the AI agent is necessary.
  • In the case of critical errors, a detailed analysis of the training data or logic may be required.

​
Conclusion

The evaluation of the test results is essential to assure the quality and reliability of the AI agents. By comparing received and expected results, weaknesses can be detected and specifically optimized.