Manual Tests
AI handles the broad coverage. Manual tests fill the gaps. And when you add tests manually, the AI enhances them with auto-generated assertions, smart selectors, and visual baselines.
GUIDE
Adding and working with manual tests
When to add tests manually
AI exploration covers a broad surface area but may miss certain scenarios — flows requiring specific test data, multi-step forms with conditional logic, email verification flows, or applications behind complex authentication. Manual test creation lets you fill these gaps.
You might also want to add tests for newly deployed features before the next scheduled exploration, replicate a customer-reported bug scenario, or test specific edge cases that require precise data inputs. Manual tests complement AI-generated tests — they don't replace them.
Creating a manual test case
From the project dashboard, navigate to Test Cases and click Create Test Case. Define the test name, the page URL it targets, the steps to execute, and the expected outcome for each step. You can reference existing flows and entities from the knowledge base.
Test steps can include actions like click, type, navigate, wait, scroll, and assert. For each step, specify the target element (by text, selector, or visual reference) and any input values. Expected outcomes can be text presence, element visibility, URL change, or custom assertions.
How AI enhances manual tests
When you create a test manually, the AI doesn't just store it — it enhances it. The AI analyzes your test steps against the knowledge base, auto-generates additional assertions, suggests selectors, fills in missing steps, and validates that the test targets real elements in your app.
For example, if you add a test for a checkout flow with a single step, the AI may suggest completing the remaining steps based on the flow structure in the knowledge base. It also attaches relevant screenshots, identifies visual regression baselines, and links the test to related issues and entities. The result is a hybrid — your intent combined with AI completeness.
Linking manual tests to the knowledge base
Manual tests are automatically linked to existing knowledge base entries when they reference the same pages, flows, or entities. This means manual tests participate in coverage analytics, appear in re-exploration results, and benefit from knowledge base updates.
When the knowledge base evolves (through re-exploration or edits), manual tests are checked for continued validity. If a referenced page or element no longer exists, you'll be notified during the next test run. This keeps manual tests as dynamic as AI-generated ones.
Updating manual tests after app changes
When your app changes, manual tests may need updates. The dashboard flags manual tests that reference elements no longer found on the page. You can edit the steps to match the new UI, or let the AI suggest updated selectors based on page content.
Manual tests also benefit from self-healing selectors — if the AI can identify the target element contextually despite the UI change, the test continues passing without intervention. Only when the AI cannot confidently resolve the change do manual tests require attention.
CAPABILITIES
How AI enhances manual test creation
Auto-generates assertions
Adds expected outcomes based on page content
Suggests missing steps
Completes partial flows from knowledge base
Validates element existence
Confirms targets are real elements on the page
Attaches screenshots
Captures visual baselines for regression detection
Links to entities and flows
Connects manual tests to knowledge base items
Self-heals selectors
Adapts to UI changes automatically
WHAT'S NEXT