- Getting started
- UiPath Conversational Agents
- UiPath Agents in Studio Web
- UiPath Coded agents
- Build with Coding Agents
Testing the agent
Test and iteratively improve an agent using the Debug configuration panel and Studio Web debugging tools.
Now it's time to test your agent and see how you can improve it.
-
In the toolbar above the Studio Web designer, select Debug. The Debug configuration window opens.
-
In the Debug configuration window, confirm the resources used in the solution and the sample input.
-
Select Save & Debug.
Note:Debug runs have a one-hour timeout. If your agent processes large files, uses many tools or tasks, or performs multiple retries, it may reach this limit and stop before completing.
-
Use the test output to refine the agent design.
To change the sample input without starting a run, select the arrow next to Debug, then select Debug configuration. This window has Save instead of Save & Debug.
Evaluating the agent
Next, go to the Evaluation sets and Evaluators panels to review and measure your agent.
- In the Evaluation sets panel, select View details on the default evaluation set. To rename the set, select the pencil icon.
- Select Add to set to add test cases with expected outcomes.
- In the Evaluators panel, review the default evaluators.
- (Optional) Add an evaluator to check anything the default evaluators don't cover.
Each evaluation links a known input with an assertion about the output.
For details, refer to Evaluations and Configuring simulations in evaluations.
Reviewing the agent [Preview]
Ask Autopilot to review your agent, or select the Review agent badge. Autopilot grades the agent and can fix the issues it finds. For details, refer to Agent review.
Using Autopilot to improve your agent [Preview]
Select Open Autopilot from the right-side panel to receive suggestions on improving prompts, tools, and other components. Such improvements support a higher agent grade and make your agent ready for deployment.