DEV Community

Rati Jha
Rati Jha

Posted on

How I Use Claude & Codex Daily as a QA Automation Engineer (Top 5 Real Use Cases)

#ai #automation #productivity #testing

AI isn’t replacing QA engineers but it is quietly replacing a lot of the repetitive, time-draining work we used to do manually.

In my day-to-day work as a QA Automation Engineer, tools like Claude and Codex have become less of a “nice-to-have” and more like a second brain. Here are the top 5 ways I actually use them in real projects, with practical examples.

  • Generating Test Specification Documents in Minutes

Writing test specs used to take hours — especially when translating requirements from Jira into structured test scenarios.

Now, I feed the requirement directly into Codex and get a clean first draft.

Example input:

`Generate a test specification for a login feature with:

  • valid login
  • invalid password
  • locked account
  • session timeout`

Output (refined):

    Test Case ID
    Preconditions
    Steps
    Expected Results
Enter fullscreen mode Exit fullscreen mode

Instead of starting from scratch, I just review and refine.

👉 Result: ~70% time saved on documentation.

  • Bulk Code Changes Without Losing My Mind

Refactoring test code across multiple files is painful — especially when patterns change.

Using Codex, I can describe the change once and apply it everywhere.

Example:

Update all Selenium locators from XPath to CSS selectors

Or:

Replace time.sleep() with explicit waits across test files

Instead of manually editing 20+ files, I:

Generate the updated pattern
Apply it across the repo after making intelligent changes instead of doing copy/paste(via IDE tools)
Enter fullscreen mode Exit fullscreen mode

👉 Result: Faster refactoring + fewer human errors.

  • Writing Python Test Scripts from Plain English

This is probably the biggest daily win.

I describe a test flow, and Codex generates a working script in Python.

Example prompt:

Write a Selenium test in Python:
1. Open login page
2. Enter username/password
3. Click login
4. Verify dashboard is visible
Enter fullscreen mode Exit fullscreen mode

Generated output (simplified):

from selenium import webdriver
from selenium.webdriver.common.by import By

driver = webdriver.Chrome()
driver.get("https://example.com/login")
driver.find_element(By.ID, "username").send_keys("testuser")
driver.find_element(By.ID, "password").send_keys("password")
driver.find_element(By.ID, "login").click()
assert "Dashboard" in driver.page_source
driver.quit()
Enter fullscreen mode Exit fullscreen mode

I still review it (always), but it removes the “blank page problem.”

👉 Result: Faster test creation, especially for repetitive flows.

  • Research & Deep Dives Using NotebookLM

When I need to understand a new tool, framework, or testing strategy, I use NotebookLM.
Become a Medium member

Instead of:

Reading 10 different blogs
Piecing together info
Enter fullscreen mode Exit fullscreen mode

I:

Upload docs / links
Ask targeted questions
Enter fullscreen mode Exit fullscreen mode

Example:

Summarize best practices for API test automation using Python

It gives:

Structured insights
Key patterns
Simplified explanations
Enter fullscreen mode Exit fullscreen mode

👉 Result: Faster learning with less noise.

  • Generating Edge Cases & Test Ideas

This one is underrated.

AI is great at thinking of scenarios you might miss.

Using Claude, I ask:

List edge cases for a payment system

Output includes:

Network failures
Duplicate transactions
Currency mismatches
Timeout scenarios
Enter fullscreen mode Exit fullscreen mode

This helps strengthen test coverage beyond “happy paths.”

👉 Result: Better quality tests with minimal extra effort.
What Still Needs Human Judgment

Let’s be real — AI isn’t perfect.

Things I never fully trust AI with:

Final test logic validation
Business-critical edge cases
Debugging flaky tests
Enter fullscreen mode Exit fullscreen mode

AI helps you move faster — but you’re still the quality gate.
Final Thoughts

The real advantage isn’t just using AI — it’s knowing where it actually saves time.

For me, that’s:

Documentation
Boilerplate code
Refactoring
Research
Idea generation
Enter fullscreen mode Exit fullscreen mode

If you’re in QA and not using AI like this yet, you’re honestly leaving a lot of efficiency on the table.

Top comments (0)