# How to Test LLM-Powered Apps: A Practical Checklist for QA Engineers (2026)

> Source: <https://dev.to/isha_kumargupta_01fbbf71/how-to-test-llm-powered-apps-a-practical-checklist-for-qa-engineers-2026-231p>
> Published: 2026-11-18 05:39:34+00:00

**Recommendation:** Replace exact-match assertions with a golden dataset of 30–100 real user questions checked for required and forbidden facts, and rerun it in CI/CD whenever the prompt, model or data changes.
More in the playbook "Evals and testing for LLM apps": <https://wpnews.pro/playbooks/evals-testing>

If you've tried testing a chatbot or an AI feature, you've probably hit this wall:

You run the same test twice and get two different answers. Both might be correct. Or both might be wrong.

Classic `assertEquals(expected, actual)` testing doesn't work well for Large Language Models (LLMs). But that doesn't mean LLM apps can't be tested. It just means we need a different checklist.

Here's the one we use when training QA engineers on AI testing.

Before writing any automation, collect **30–100 real questions** your users will ask, along with what a *good* answer must contain (not the exact wording).

| Question | Must include | Must NOT include | 
|---|---|---|
| What is your refund policy? | "7 days", "full refund" | Made-up phone numbers | 
| Can I reset my password? | Link to reset page | Asking for current password | 

This becomes your regression suite. Every time the prompt, model or data changes, you rerun it.

Instead of checking the exact sentence, check for **key facts** or use a second model as a judge.

``` python
def test_refund_policy():
    answer = call_llm("What is your refund policy?")

    # Check key facts instead of exact wording
    assert "7 days" in answer
    assert "refund" in answer.lower()

    # Check what should NOT be there
    assert "call us at" not in answer.lower()
```

For more advanced checks, tools like **DeepEval**, **Ragas** and **Promptfoo** can score answers for relevance, correctness and faithfulness.

If your app uses RAG (answers from your documents), every claim in the answer should be traceable to the source documents.

Test it by asking questions whose answers are **not** in your documents. A good system says *"I don't know"*. A bad one invents an answer.

Your users (or attackers) will try things like:

Keep a list of these attack prompts and run them in every release. This is the **security testing** of the AI world.

Run the same question **5–10 times**. The wording can change, but the facts shouldn't. If one run says "7 days" and another says "30 days", you have a reliability bug.

LLM apps have new performance metrics:

Your existing skills with JMeter, k6 or Playwright still apply here.

Check that the app:

AI providers update models regularly. A prompt that worked perfectly last month might behave differently today. **Automate your golden dataset** and run it in CI/CD, just like any regression suite.

AI testing is quickly becoming one of the most in-demand QA skills. At [Isha Training Solutions](https://ishatrainingsolutions.org) we've been training working professionals for about 10 years, and our **AI & LLM Testing** program covers all of the above hands-on, along with Playwright, JMeter and Claude-assisted test design.

👉 Explore the courses: [ishatrainingsolutions.org](https://ishatrainingsolutions.org)

What's the hardest part of testing AI features in your project? Share in the comments. I'd love to hear how other teams are handling it! 👇
