Märkus.
Juurdepääs sellele lehele nõuab autoriseerimist. Võite proovida sisse logida või kausta vahetada.
Juurdepääs sellele lehele nõuab autoriseerimist. Võite proovida kausta vahetada.
[This article is prerelease documentation and is subject to change.]
After you run an evaluation, review the results to see how well your agent performed across conversations. The results show quality scores and let you open individual conversations to see the agent's responses.
Important
- This is a production-ready preview feature.
- Production-ready previews are subject to supplemental terms of use.
View run results
To view run results:
- Open your agent in Copilot Studio.
- Select the Evaluate tab.
- Under recent results, select an evaluation with a prior test run to view its results.
The results show the following information:
Test run result: A table that shows each conversation with the number of messages between the user and the agent, and the general quality score. Select a conversation to see more details.
Evaluation summary: A summary of your overall evaluation across all test methods. It includes the following information:
- Score: The overall score of conversations that passed the general quality test method.
- Duration: The length of time it took to complete the evaluation.
- Test cases: The number of test cases in your overall evaluation.
- Data type: The type of test set. Only the Conversation data type is available.
- User profile: The user profile that ran the evaluation.
Review a test case
To review a test case:
On the run results page, select a conversation to expand its details.
Review the Test case details:
- User messages: The test messages sent to the agent.
- Agent responses: The agent's actual responses.
- General quality: How the responses scored on the test method, either Pass or Fail.
Use this information to identify where your agent needs improvement.
Compare runs
Run the same evaluation multiple times to compare results and track progress:
- On the evaluation detail page, review the list of runs with their scores and timestamps.
- Compare scores across runs to see if your agent changes improved or degraded performance.
- Look for patterns. For example, if a category of conversations consistently scores low, that area might need more instructions or knowledge.
Export results
Export evaluation results for further analysis:
- On the run results page, select the three dots (…) > Export test results. You can also select a specific evaluation, and then select the three dots (…) > Export test results in the upper right corner.
- Download the results as a CSV file that includes all conversations, responses, and scores.
Act on results
Based on your evaluation results:
- Low general quality scores: Review and refine your agent's instructions. See Configure agent details and instructions (preview).
- Missing or incorrect tool use: Check that tools have clear descriptions and that your agent's instructions say when to use them. See Tools overview for agents (preview).
- Incorrect information: Verify that the relevant knowledge sources are added and configured correctly. See Add knowledge to an agent (preview).