Test with AI
Generative AI is not going to replace testers. Testers who use AI well are going to replace testers who do not. This module teaches you the difference.
1 Why Test with AI
Every test team in Aotearoa is being asked the same question right now: "how are you using AI?" Most testers answer with either fear ("it will take my job") or bluster ("I use ChatGPT to write test cases"). Both answers are wrong.
The honest answer is that generative AI has become a genuine productivity multiplier for testers who learn to prompt it deliberately, validate its output rigorously, and integrate it safely into a test process. It can triple your test case coverage on a user story. It can spot defect patterns a human would miss. It can turn a vague bug report into a reproducible scenario in thirty seconds.
It can also hallucinate test data that looks real but violates your spec. It can leak confidential requirements to a third party. It can give you false confidence in a broken test suite. Testers who do not understand the risks will be the testers who get replaced — not by the AI, but by testers who do understand the risks.
This module teaches you to do it properly.
2 Mapped to ISTQB CT-GenAI
The ISTQB Certified Tester — Testing with Generative AI (CT-GenAI) is the official international certification for this skill set. It was released on 25 July 2025 — very new territory. No other NZ bootcamp covers it yet.
The certification covers 5 chapters, 24 learning objectives, and a minimum of 13.6 hours of instruction. Entry requirement is ISTQB Foundation Level (CTFL). The exam is 40 multiple-choice questions, 60 minutes, with a 65% pass mark.
This module is structured to prepare you for every learning objective in the CT-GenAI syllabus, using the same pedagogy as the rest of Resync Bootcamp: worked examples, NZ business context, common mistakes, and hands-on practice.
3 The 5 Modules
1 GenAI Foundations for Testers
The spectrum of AI: symbolic AI, classical machine learning, deep learning, and generative AI. How large language models work under the hood: tokenisation, context windows, embeddings. The difference between foundation, instruction-tuned, and reasoning LLMs. Multimodal models that read images. What LLMs can and cannot do for testers.
CT-GenAI Ch 1 · GenAI-1.1.1 to 1.2.2
2 Prompt Engineering for Testing
How to structure a prompt: role, context, instruction, input data, constraints, output format. Zero-shot, one-shot, few-shot, prompt chaining, meta-prompting. System prompts vs user prompts. Apply GenAI to test analysis (refining acceptance criteria for a Kiwisaver signup story), test design (generating Gherkin scenarios for IRD validation), regression (keyword-driven scripts), and monitoring. Evaluate and iteratively refine your prompts.
CT-GenAI Ch 2 · GenAI-2.1.1 to 2.3.2
3 Managing the Risks of AI in Testing
Hallucinations, reasoning errors, and bias — how to spot them in LLM output and mitigate them. Data privacy and security when feeding test artefacts to a third-party model (critical under the NZ Privacy Act 2020). Non-determinism and how to test around it. Energy consumption and environmental impact. AI regulations, standards, and best practice frameworks.
CT-GenAI Ch 3 · GenAI-3.1.1 to 3.4.1
4 LLM-Powered Test Infrastructure
Architectural components of an LLM-powered test tool. Retrieval-Augmented Generation (RAG) — grounding AI output in your own API documentation or test spec. LLM-powered agents that can automate multi-step test processes. Fine-tuning language models for specific test tasks. LLMOps for deploying and managing LLMs in a test organisation.
CT-GenAI Ch 4 · GenAI-4.1.1 to 4.2.2
5 Adopting GenAI in Your Test Organisation
Shadow AI and its risks (unapproved tools leaking IP and breaching privacy). Defining a GenAI test strategy. Selecting the right LLM or small language model for a given test task and context. Estimating recurring costs. Building AI capability in a test team. How test processes and tester responsibilities shift when GenAI is adopted.
CT-GenAI Ch 5 · GenAI-5.1.1 to 5.2.3
4 Build Progress
This section is in active development. Here is what is being built:
- 5 module pages mirroring the CT-GenAI chapter structure
- NZ-specific prompt template library (IRD validation, Kiwisaver flows, ANZ/BNZ API specs, RealMe auth, EQC claim forms)
- Live hands-on exercises backed by a real LLM — you will actually prompt a model and evaluate the output
- Worked examples with planted flaws for hallucination-detection practice
- Self-check questions mapped to every CT-GenAI learning objective
- Study guide PDF for the CT-GenAI exam
Check back soon. If you would like early access or to contribute NZ testing scenarios, get in touch via the main Resync site.