Before making a Zoho Zia Agent available to users, it is important to validate that it behaves as intended across a range of scenarios including edge cases. Zoho provides a built-in testing environment directly within Agent Studio for this purpose.

Connection Prerequisites

Any tool that connects to an external service — such as Zoho CRM or Zoho Desk — requires an established service connection before testing can begin. You can set these up in two ways:

• Via the agent's Tools section using the "Manage Connections" option.
• Via Settings > Connections, which allows connections to be configured in advance before building the agent.

Tool Parameter Sources

Before running tests, you need to specify how each tool parameter receives its value. Three source types are available:

• Model-driven. The agent determines the parameter value autonomously, based on its instructions and the conversation history to that point.
• On-Trigger. The value is supplied dynamically at the point the agent is activated. During testing, these parameters must be entered manually in the Edit Parameters section.
• Constant. A fixed, predefined value that is used consistently every time the tool runs, regardless of context.

Optional parameters can be disabled for test scenarios where they are not relevant.

Testing Workflow

Work through the following steps to run a test session:

1. Open the test environment. Navigate to the agent's details page and click the "Test Agent" button.
2. Configure service connections. Ensure all required connections for the tools you intend to test are active.
3. Open the Tools Param Mapping panel. This panel displays each tool and its parameters so you can specify the source type for each one.
4. Designate source types. Set each parameter to Model-driven, On-Trigger, or Constant as appropriate for your test scenario.
5. Enter On-Trigger values. In the Edit Parameters section, provide values for any parameters set to On-Trigger.
6. Submit test prompts. Enter prompts that reflect realistic user queries aligned with the agent's configured tools and defined role.
7. Evaluate responses. Check that the agent's answers are accurate, appropriately sourced, and within the scope defined by your instructions and guardrails.
8. Clear and repeat. Use the "Clear Chat" button to reset the conversation history between distinct test scenarios.
9. Iterate until reliable. Continue refining instructions, guardrails, and tool configurations until the agent responds consistently across all test cases.

Exporting Parameters for API Deployment

The "Copy Parameters" button within the testing panel exports the current tool configuration in a format compatible with API deployment requests. This is useful when you intend to deploy the agent programmatically and want to replicate the tested configuration exactly.