Managing Tests with Your AI

You created a test. It worked. You created another. Then ten more. Now you have 50 tests and you're not sure which ones actually matter, which ones are broken, or which ones check features you deleted three months ago.

Opening each test manually to see what it does takes forever. Deleting them one by one is worse. You need a way to work with your tests that doesn't require clicking through a UI for an hour.

Your AI can do all of this. You just tell it what you want in plain language.

I Deleted Tests by Mistake

You meant to delete one test. You deleted five. Or you cleaned up old tests and realized you needed one of them.

I just deleted some tests by mistake. Undo that.

I'll restore your last deletion. [Status: Undoing update eeb3a465-22c1-485b-bb3a-419703113a31]

Done. Restored 5 tests:

  • Homepage Load
  • Checkout Flow
  • Login Test
  • Product Search
  • Cart Update

Your tests are back. This works for any update you made recently—deletions, renames, content changes, tag updates.

I Have Old Tests Cluttering Everything

You redesigned checkout three months ago. The old checkout tests are still running, failing constantly, and making it hard to see what's actually broken. You need them gone.

Delete all tests tagged #oldCheckout

I'll delete all tests with the oldCheckout tag. [Status: Finding tests with tag oldCheckout] [Status: Found 23 tests]

Deleted 23 tests tagged #oldCheckout. If you need them back, tell me to undo this update.

Twenty-three tests gone. Dashboard is clean. You can still undo it if you deleted the wrong ones.

Which Tests Are Actually Broken?

Something's failing. You got an alert. You open the dashboard and see 50 tests. You have no idea which ones failed or when.

Show me all failing tests.

Here are your failing tests: Currently Failing (2 tests):

  • Checkout Flow - Last failed 5 minutes ago Error: Button Complete Purchase not found
  • Login Page Load - Last failed 2 hours ago Error: Page returned 500 status Recently Fixed (1 test):
  • Product Search - Was failing, now passing (fixed 10 minutes ago)

Now you know exactly what's broken and what just got fixed. You can focus on the two that actually need attention.

Run Tests Before Deploying

You just changed the checkout flow. You want to make sure it still works before you push to production. You have 15 checkout-related tests and don't want to run all 50 tests.

Run all #checkout tests.

I'll run all tests tagged checkout. [Status: Running 15 tests with tag checkout] [Status: 14 passed, 1 failed]

Results:

  • 14 tests passed
  • 1 test failed: Payment Form Validation Error: Field Card Number not found The payment form test is failing. Check if you renamed the card number field.

You caught the break before deploying. The card number field changed and the test needs updating. Fix it now instead of finding out from customers.

What Does This Test Actually Do?

A test called "User Flow Test 3" is failing. You have no idea what it checks. You don't want to open it in the UI and read through 20 lines of commands.

Show me what User Flow Test 3 does.

Test: User Flow Test 3 What it does:

  1. Goes to /products
  2. Clicks Filter by Price
  3. Selects $0-$50 range
  4. Verifies at least 5 products are shown
  5. Clicks first product
  6. Checks product detail page loads Currently: Failing on step 4 - only 2 products shown Tags: #products, #filters, priority:high Last run: 3 minutes ago

Now you know what it does and exactly where it's failing. The filter isn't returning enough products—probably a data issue.

Organize Tests with Tags

You have 100 tests. Some have tags like "priority:critical" and "feature:checkout". Most have no tags at all. You can't find anything.

Add tag #payment to all tests that mention checkout or purchase.

I'll tag all checkout-related tests with #payment. [Status: Searching test content for checkout and purchase] [Status: Found 18 tests]

Added #payment tag to 18 tests:

  • Checkout Flow
  • Complete Purchase
  • Payment Form Load
  • Card Validation ... (14 more tests)

Now you can find all payment-related tests by filtering for #payment. Your dashboard is organized without manually editing 18 tests.

Fix a Test After Code Changed

You changed a button from "Buy Now" to "Purchase". The test is failing because it's looking for the old button text.

Fix the Add to Cart test - the button text changed from Buy Now to Purchase.

I'll update the Add to Cart test with the new button text. [Status: Updating test content]

Updated Add to Cart test:

  • Old: Click text=Buy Now
  • New: Click text=Purchase Running test to verify fix... [Status: Test passed] Test is now passing with the new button text.

Test fixed and verified. You didn't write any code or open any editor.

Rename Tests After Feature Renamed

You renamed the "Products" section to "Catalog". You have 12 tests with "Products" in the name. You need them all renamed to match.

Rename all tests - change Products to Catalog.

I'll rename all tests that mention Products. [Status: Finding tests with Products in name] [Status: Found 12 tests]

Renamed 12 tests:

  • Products Page Load → Catalog Page Load
  • Products Search → Catalog Search
  • Products Filter → Catalog Filter ... (9 more tests)

All tests renamed. Your test names match your actual feature names again.

Create Similar Tests

You have a "Homepage Load" test that checks if the homepage loads. You need the same test for /pricing, /about, and /contact pages.

Create tests like Homepage Load but for the /pricing, /about, and /contact pages.

I'll create three new uptime tests based on Homepage Load. [Status: Creating Pricing Page Load test] [Status: Creating About Page Load test] [Status: Creating Contact Page Load test]

Created 3 new tests:

  • Pricing Page Load - checks /pricing loads
  • About Page Load - checks /about loads
  • Contact Page Load - checks /contact loads All tests use the same tags and priority as Homepage Load.

You went from one test to four tests with one sentence. No copying, no pasting, no editing.

Find Tests by What They Check

You changed how login works. You need to find every test that uses login so you can verify they still work. You're not sure which tests actually log in.

Show me all tests that check login.

Found 8 tests that involve login: Tests that navigate to /login:

  • Login Page Load
  • Login Form Validation
  • Login Error Handling Tests that use As Logged In User state:
  • Dashboard Access
  • User Profile Page
  • Settings Update
  • Account Deletion
  • Payment Methods All of these tests depend on login functionality working correctly.

Now you know exactly which tests to check after changing login. Eight tests instead of manually reviewing all 100.

What This Changes

Before: Managing tests means opening each one, reading through code, manually updating names and tags, clicking through UI menus, hoping you didn't miss anything.

After: Managing tests means telling your AI what you want in plain language. It finds, updates, organizes, runs, and fixes tests while you work on actual problems.

You're not learning new tools or memorizing commands. You're just describing what you need, and your AI handles the details.