Back to index
OtherTypically a 2-day accredited course or self-study; 60-minute exam (plus 25% for non-native speakers)·Set by local exam providers; UK two-day courses with the exam sit in the high hundreds to low thousands of pounds

ISTQB Certified Tester Testing with Generative AI (CT-GenAI)

3.9

This one is about testers using LLMs to do their job better, not about testing AI products. That makes it more immediately useful than CT-AI for most people, and the prompt engineering and risk chapters are practical. It is still a short multiple choice exam, so treat it as a structured way to learn the ground rules rather than proof of skill.

What We Liked

  • Focused on a problem testers have right now: using LLMs for test design, test data and automation
  • Risk chapter is sensible about hallucinations, data leakage and over-trust in generated tests
  • Short and focused: five chapters, 40 questions, 60 minutes
  • v1.1 update in 2026 added context on LLM-powered agents without changing the exam scope
  • Free syllabus and official sample exam make self-study easy

What Could Be Better

  • Needs CTFL first, same as every ISTQB specialist module
  • Tool-agnostic by design, so you will not learn any specific product in depth
  • Content dates fast; agentic testing tools have moved a lot since the 2025 syllabus
  • No official price list, and classroom bundles are expensive
  • Easy to confuse with CT-AI, which covers testing AI systems rather than testing with AI

Detailed review

CT-GenAI is the newer of the two ISTQB AI certifications and the one I would point most testers to first. Where CT-AI is about testing machine learning systems, this one is about using generative AI as a tester. The syllabus v1.0 is dated July 2025 and ISTQB announced a minor v1.1 update on 27 April 2026. The update tidies terminology, risk content and hands-on objectives, and adds context on LLM-powered agents, but ISTQB is clear it does not change the structure, objectives or scope, so a v1.0 course is not wasted. The exam is 40 questions worth 46 points, pass mark 30 (65%), in 60 minutes plus 25% extra for non-native speakers.

CTFL is a prerequisite. The five chapters are an introduction to GenAI for software testing, prompt engineering for testing, managing GenAI risks in the test process, LLM-powered solutions for testing, and deploying and integrating GenAI in a test organisation. The prompt engineering and risk chapters are where I think the value is. Most teams I see are using LLMs to draft test cases with no review process, then wondering why coverage looks great on paper and the bugs still ship. A syllabus that makes people think about verification of generated output, data privacy when pasting production data into a chatbot, and non-deterministic results is useful even if the exam itself is basic. The last chapter, on rolling GenAI out across a test organisation, is aimed at leads and managers.

If you are an individual contributor it will feel abstract. Search volume is tiny right now, around 10 a month for the GenAI terms in the US, but that mostly reflects how new it is. I expect it to overtake CT-AI over the next year or two because more testers use LLMs than test them. Price is the same story as every ISTQB module. Exam fees are set locally and UK providers sell two-day virtual courses with the exam bundled. With a free syllabus and a published sample exam, self-study is the sensible route unless your employer is paying for the classroom.

[ final ]

The verdict.

Worth it for test leads and QA managers who need a shared vocabulary for bringing GenAI into a test team, and for consultancies who need a badge for it. Individual testers who already use Copilot or Claude every day for test work will learn less than they expect. Self-study it and skip the course if your employer is not paying.