Skip to main content

​
Overview

AI Test Generation allows you to create complete, ready-to-run test cases by describing your testing goals in natural language. The AI analyzes your description and generates appropriate test steps, including navigation, interactions, and assertions.

​
How It Works

  1. Describe your test: Write what you want to test in plain English
  2. AI analyzes: The system interprets your intent and identifies required steps
  3. Steps generated: Complete test steps are created with appropriate locators
  4. Review and refine: Edit generated steps as needed before running

​
Getting Started

​
From the Dashboard

  1. Click New Test or use the + button
  2. Select Generate with AI
  3. Enter your test description
  4. Click Generate
  5. Review and save the generated test

​
From AI Chat

  1. Open the global AI Chat
  2. Describe the test you want to create
  3. The AI generates test steps
  4. Save as a new test case

​
Writing Effective Descriptions

​
Structure Your Description

Include these elements for best results:
  • Starting point: Where the test begins
  • Actions: What steps to perform
  • Verifications: What to check or assert
  • Test data: Specific values to use

​
Example Descriptions

Basic Login Test: 
Test user login on myapp.com:
- Navigate to the login page
- Enter email "[email protected]"
- Enter password "password123"
- Click the Sign In button
- Verify the dashboard page loads
- Verify "Welcome back" message is displayed
Form Submission Test: 
Test the contact form submission:
- Go to /contact
- Fill in name field with "John Doe"
- Fill in email with "[email protected]"
- Fill in message with "Test inquiry message"
- Click Submit
- Verify success message "Message sent successfully" appears
E-commerce Flow: 
Test adding items to cart on an e-commerce site:
1. Navigate to the products page
2. Click on the first available product
3. Select quantity 2
4. Click Add to Cart
5. Verify cart badge shows "2"
6. Go to cart page
7. Verify product appears in cart with correct quantity

​
Generated Step Types

AI Test Generation creates various step types based on your description:
Description ElementGenerated Step Type
”Go to”, “Navigate to”Navigate
”Click”, “Press”, “Select”Click
”Enter”, “Type”, “Fill”Fill
”Verify”, “Check”, “Assert”Verify Text/Visibility
”Wait for”Wait for Navigation/Element
”Upload”File Upload
”Select option”Select Option

​
Customizing Generated Tests

​
Adding Environment Variables

Generated tests use hardcoded values by default. Replace with environment variables: Before: 
Navigate to: https://myapp.com/login
After: 
Navigate to: {{ env.BASE_URL }}/login

​
Improving Locators

AI generates locators based on common patterns. Improve them for stability:
  • Use data-testid attributes when available
  • Prefer role-based selectors
  • Add fallback locators for reliability

​
Adding Assertions

Enhance generated tests with additional verifications:
  • Add visual assertions for complex layouts
  • Include URL verification after navigation
  • Add wait steps for dynamic content

​
Use Cases

​
Rapid Prototyping

Quickly create test skeletons for new features: 
Test the new checkout flow:
- Add product to cart
- Proceed to checkout
- Fill shipping information
- Select payment method
- Complete purchase
- Verify order confirmation

​
Coverage Expansion

Generate tests for common scenarios: 
Test password reset flow:
- Click forgot password link
- Enter email address
- Submit form
- Verify confirmation message
- Check email for reset link

​
Regression Test Creation

Build regression suites from feature descriptions: 
Test user profile editing:
- Navigate to profile settings
- Update display name
- Change email address
- Update avatar image
- Save changes
- Verify all changes are reflected

​
Best Practices

  1. Start with clear goals: Know what you want to verify before describing
  2. Include specific values: Use actual test data in descriptions
  3. Describe expected outcomes: Tell AI what success looks like
  4. Review all locators: Verify selectors match your application
  5. Add error handling: Include steps for expected error states
  6. Use environment variables: Replace hardcoded URLs and credentials

​
AI Credit Usage

ActionCredits
Generate test from description1 credit
Credits are consumed when the generation process starts, regardless of outcome.

​
Limitations

  • Cannot access your live application during generation
  • May not know application-specific selectors
  • Complex conditional logic requires manual setup
  • Multi-tab scenarios need manual configuration

​
Troubleshooting

​
Generated Steps Don’t Match App

  • Provide more specific descriptions
  • Include actual selector hints if known
  • Use the recorder to capture accurate locators

​
Missing Assertions

  • Explicitly state what to verify in your description
  • Add verification phrases like “verify”, “check”, “assert”

​
Incorrect Navigation

  • Specify full URLs or paths
  • Describe the navigation flow step by step