
Preface
The International Software Testing Qualifications Board (ISTQB®) officially released the Expert Level Certification Syllabus - Testing with Generative AI (CT-GenAI) v1.0 on July 25, 2025. This is the world’s first professional certification standard focused on the application of generative AI in the field of software testing. It aims to provide testing professionals with a systematic body of knowledge, helping them effectively use generative AI technologies to improve testing efficiency and quality.
The syllabus authors are: Abbas Ahmad, Gualtiero Bazzana, Alessandro Collino, Olivier Denoo, and Bruno Legeard
The following content is organized based on the syllabus and can be used as a reference.
Certification Background and Objectives
Core Value Positioning
The CT-GenAI certification is intended for all professionals involved in software testing, including:
- Test engineers, test analysts, test automation engineers
- Test managers, user acceptance testers, software developers
- Project managers, quality managers, software development managers
- Business analysts, IT directors, and management consultants
Five Key Business Outcomes
Professionals who obtain this certification will have the following core capabilities:
- Understand the basic concepts, capabilities, and limitations of generative AI
- Master large language model prompt engineering skills for software testing
- Identify and mitigate the risks of generative AI in software testing
- Understand the applications of generative AI solutions in software testing
- Effectively participate in the formulation and implementation of generative AI testing strategies within the organization
Certification Content System
Chapter 1: Fundamentals of Generative AI in Software Testing
Core Concepts of Generative AI
- AI technology lineage: The differences and application scenarios of symbolic AI, classical machine learning, deep learning, and generative AI
- Large Language Model (LLM) fundamentals:
- Tokenization and Embedding technologies
- The concept of Context Window and its impact
- Non-deterministic behavior characteristics
- LLM classification system:
- Foundation LLMs
- Instruction-Tuned LLMs
- Reasoning LLMs
- Multimodal LLMs and Vision-Language Models: The ability to process multimodal data such as text, images, and audio
Core Capabilities of Generative AI in Testing
- Requirements analysis and improvement: Identify ambiguity, inconsistency, and missing information in requirements
- Test case generation: Create test cases based on system requirements or user stories
- Test oracle generation: Automatically generate expected results
- Test data generation: Create datasets, boundary values, and test data combinations
- Test automation support: Generate test scripts from test case descriptions
- Test result analysis: Create summaries and classify anomalies by severity and priority
- Test artifact creation: Generate test plans, test reports, and defect reports
Comparison of Interaction Modes
- AI chatbots: Provide a natural-language interaction interface, suitable for quick feedback and concept clarification
- LLM-driven testing applications: Enable automated testing tasks through API integration, supporting customization and scalability
Chapter 2: Prompt Engineering for Effective Software Testing
Effective Prompt Generation
- Six elements of prompt structure:
- Role: Define the perspective the AI should adopt (e.g., tester, test manager)
- Context: Provide the test object and background information
- Instruction: Specify the task to be performed
- Input Data: Provide specific test materials
- Constraints: Set limits and rules
- Output Format: Specify the desired response format
Core Prompting Techniques
- Zero-shot Prompting: Request a task directly without examples
- Few-shot Prompting: Provide a small number of examples to guide the AI
- Prompt Chaining: Break a complex task down into multiple sequential prompts
- Meta Prompting: Guide the AI on how to think about the task itself
Test Task Applications
- Test analysis: Use multimodal prompts to generate acceptance criteria for GUI wireframes
- Test design and implementation:
- Generate functional test cases from user stories
- Use few-shot prompting to generate Gherkin-style test conditions
- Determine test case priorities through prompt chaining
- Automated regression testing:
- Create and manage keyword-driven test scripts
- Analyze test reports with structured prompts
- Test monitoring and control: Generate monitoring metrics based on test data
Chapter 3: Managing the Risks of Generative AI in Software Testing
Key Risk Types
- Hallucinations and reasoning errors:
- Identify factual errors and logical flaws in LLM output
- Mitigation techniques: fact-checking, multi-model verification, human review
- Data privacy and security risks:
- Privacy vulnerabilities in test data and tools
- Mitigation strategies: data masking, local deployment, access control
- Energy consumption and environmental impact:
- Assess the energy consumption and carbon emissions of GenAI usage
- Optimize model usage efficiency and reduce resource waste
Compliance Framework
- AI regulations and standards: EU AI Act, ISO/IEC 38507, etc.
- Best practice frameworks: Responsible AI usage guidelines, ethical testing principles
Chapter 4: LLM-Driven Testing Infrastructure
Architectural Approaches
- Key architectural components:
- Model service layer
- Prompt engineering layer
- Result evaluation layer
- Integration interface layer
- Retrieval-Augmented Generation (RAG): Combine external knowledge bases to improve response accuracy
- LLM-driven testing agents: Intelligent agents that automate the testing process
Model Optimization and Operations
- Fine-Tuning: Optimize LLMs for specific testing tasks
- LLMOps (Large Language Model Operations):
- Model deployment and version management
- Performance monitoring and optimization
- Cost control strategies
Chapter 5: Deployment and Integration of Generative AI in Testing Organizations
Adoption Roadmap
- Shadow AI risks: Security risks brought by the use of unapproved AI tools
- Key strategic elements:
- Define clear business goals and value propositions
- Select suitable LLMs/SLMs (Small Language Models)
- Phased implementation plan
- Adoption stages:
- Pilot and proof of concept
- Technical capability building
- Process integration
- Scale-up and promotion
Change Management
- Core skills and knowledge:
- Prompt engineering skills
- AI ethics and risk awareness
- Test automation and AI integration skills
- Team capability building:
- Develop training plans
- Build communities of practice
- Bring in external experts
- Evolution of testing processes:
- Integration of traditional testing processes with AI-enhanced processes
- Define new roles and responsibilities
- Continuous improvement mechanisms
Certification System and Exam Requirements
Cognitive Level Classification
- K1 (Remember): Identify and recall terms and concepts
- K2 (Understand): Explain and compare concepts and relationships
- K3 (Apply): Apply knowledge and skills in specific situations
Practical Objective Levels
- H0: Demonstration observation (not performed by the learner)
- H1: Guided practice (the learner follows instructions)
- H2: Practice with prompts (completed independently by the learner)
Exam Requirements
- Prerequisite: Hold an ISTQB® Foundation Level certificate
- Exam content: Based on all chapters of this syllabus (except the introduction and appendices)
- Exam format: Multiple-choice questions requiring the integrated application of knowledge from multiple chapters
Certification Value and Applicable Scenarios
Career Development Path
The CT-GenAI certification is an important part of the ISTQB® Expert Level certification system, and holders can continue to advance:
- Advanced Level certifications: Test Analyst, Technical Test Analyst, Test Manager, Test Engineer
- Expert Level certifications: Test Management, Test Process Improvement
Industry Application Value
- Efficiency improvement: Reduce repetitive testing work through AI automation
- Quality enhancement: Use AI analytical capabilities to find deeper defects
- Cost optimization: Allocate testing resources reasonably and reduce overall testing costs
- Innovation drive: Explore new application scenarios of AI in the testing field
- Risk management: Systematically identify and mitigate risks related to AI testing
Conclusion
As generative AI technology continues to evolve, testing professionals need to continuously update their knowledge systems. While enjoying the efficiency gains brought by AI, they must maintain a deep understanding of the essence of testing — ensuring the quality and reliability of software products.
Learning about the CT-GenAI certification can serve as a reference for a systematic body of knowledge and a practical guide.
For certification details, you can refer to the official ISTQB website in China: CSTQB
Appendix
https://isqi.org/media/ce/63/da/1753702121/CT-GenAI%20-%20Syllabus%20v1.0_all_.pdf
