Test Suites

A test suite is a collection of test cases grouped together for a specific agent. Suites let you run all your test scenarios at once and track results over time.

Accessing Test Suites

Navigate to Testing in the sidebar to see your list of test suites.

Test Suite List

The test suite list displays:

Click any suite to view its details, manage its test cases, or run it.

Creating a Test Suite

1. Go to Testing in the sidebar
2. Click Create Test Suite
3. Fill in the details:
   • Name — A descriptive name (e.g., “Sales Qualifier - Core Scenarios”)
   • Description — What this suite covers and when to run it
   • Agent — Select the agent this suite will test
4. Click Create

The suite is created empty. Next, add test cases to define the scenarios you want to validate.

Editing a Test Suite

To modify a suite’s name, description, or agent assignment:

1. Open the test suite
2. Click Edit in the top right
3. Make your changes
4. Click Save

Deleting a Test Suite

1. Open the test suite
2. Click Edit in the top right
3. Click Delete Suite
4. Confirm the deletion

Deleting a suite removes all its test cases and run history.

Running a Test Suite

To execute all test cases in a suite:

1. Open the test suite
2. Click Run Tests
3. The suite begins executing each test case sequentially

See Test Runs for details on how runs work and how to interpret results.

Best Practices

Organize by Agent

Create one primary test suite per agent. Name it clearly so your team knows which agent it covers.

Cover the Critical Path

Start with test cases for the scenarios your agent encounters most often. A sales qualification agent should have cases for interested callers, uninterested callers, and callers with questions about pricing.

Add Edge Cases Over Time

After the core scenarios are covered, add cases for edge cases: callers who are rude, callers who ask off-topic questions, callers who speak very little, and callers who try to break the conversation flow.

Run After Every Prompt Change

Any time you modify an agent’s prompt, re-run the test suite to catch regressions. A change that improves one scenario may break another.