Path for testers and QA engineers
Who tests what AI wrote?
Four stages from using AI in daily test work to testing AI features for risk.
- 4 stages
- 144 hours
- 18 weeks
Your pace, hours a week
4 stages, 144 hours
Stage 1, AI in daily test work
Mark its homework
This week
Have AI draft test cases for one requirement, then mark what is wrong or missing.
Paid options
Ranked by fit, ratings and freshness. We earn nothing from these links.
You end up with
AI-drafted cases for a sample requirement, marked up by you.
Answers these problems
More about this path
Where this gets you
Can review and repair AI-generated tests, build an eval set for an AI feature, and run a risk-based test of an LLM feature including prompt injection.
Just lost your job?
If you have just lost your job, do this first
Do the Google module and CT-GenAI chapters 1 and 2, about 8 hours of syllabus. Then make one public worked example: AI-drafted test cases for a sample requirement, marked up with what was wrong or missing. If postings you want require ISTQB Foundation, it can move earlier. Check the postings first.
What you need
If you are new to testing, read the free ISTQB Foundation syllabus first, linked under the exams below. Stage 2 needs some willingness to write code.
Worth the fee, or not
| Certificate | Fee | Valid | Verdict |
|---|---|---|---|
| ISTQB Foundation (CTFL) | $229 | No expiry | The only one with a real case, since other ISTQB exams require it. No hiring data found. |
| CT-GenAI | $199 | No expiry | Using AI to test. New, with no hiring evidence found. Use the free syllabus as a study map. |
| CT-AI v2.0 | $199 | No expiry | Testing AI systems. About two thirds of the syllabus is machine learning that this path does not teach. No hiring evidence found. Check you are getting version 2. |
US prices, and ISTQB says fees are set locally. Foundation plus one AI exam is $428, a real cost without income, so it comes last.
What the data says
UK manual-testing share of permanent IT ads fell from 0.31% in 2024 to 0.17% in 2026, but the number of ads did not fall: 206, then 129, then 215. Nothing proves AI caused this. Budget cuts and shifting testing left are live alternatives.
IT Jobs Watch, 21 Nov 2026. Weak to moderate: the share fell while the count of ads did not.
US QA and tester jobs are projected up 6% from 2025 to 2035, against 10% for developers. We found no reliable dataset that counts QA layoffs.
US Bureau of Labor Statistics, 27 Aug 2026. Moderate to strong: official, but a projection.
46% have shipped AI code that later failed in production.
SmartBear, 30 Sep 2026. Weak to moderate: SmartBear sells testing tools.
What this path cannot do
- Evaluation is a real skill and an unproven door for testers. Very few of those roles are junior.
- Finishing this path builds skill and gives you things to talk about. It cannot promise work.
Checked 21 November 2026.
