Welcome to the “AI in Testing” series — a collection of posts based on training material I prepared from hands-on experience with GenAI models: Claude Opus, ChatGPT, and Gemini Pro.
Who is this series for?
For testers — manual and automation — who want to understand how to practically use AI in everyday work. No theory, no marketing. Concrete frameworks, tools, and scenarios.
What will you find in the upcoming posts?
- Prompt frameworks — from simple (RTF, ARC) to advanced agent-level (ReAct+, P2P2). Each with a QA example.
- AI tools review — ChatGPT, Claude, Copilot, Codex, Rovo, Devin, Gemini, OpenClaw. What for what and when.
- Model Context Protocol (MCP) — a standard connecting AI with Jira, Confluence, Slack, and other tools.
- Practical examples — ready-to-copy prompts: test cases, bug descriptions, business cases, test code.
- Multi-agent — how to orchestrate multiple AI tools together in real QA scenarios.
- CLAUDE.md / AGENTS.md — how to configure an AI agent in your repository.
Difficulty level legend
Throughout the series I use the following labels:
- 🟢 Easy — good starting point, no AI experience required
- 🟡 Medium — requires understanding of context and precision
- 🔴 Advanced — requires experience with agents or complex prompt engineering
Sources
I developed this material based on my own experience, using Claude Opus 4.6, ChatGPT 5.4, and Gemini 3.1 Pro — synthesized and verified. Review: Konrad Gomulski, Rafal Walecki.
Join me in the next post where we start with the two most important frameworks: CRISPE and Chain of Thought.