Accelerate Testing with Generative AI (Practical Training) [EN]

Throughout this training, you will explore the fundamentals of AI for testing, adopt best practices, and discover tools to optimize your testing processes. You will develop key skills to create relevant prompts and manage error risks associated with AI models. Learn to master conversational AI to query various LLMs effectively and integrate them into your tool chains. Through practical exercises, you will have the opportunity to experiment with a variety of AI models and tools for testing. Optimize your testing processes and enhance the quality of your applications by leveraging generative AI. By the end of the training, you will be able to identify the opportunities, limitations, and risks associated with the use of generative AI for testing.

Content

1. Introduction to Generative AI for software testing

• Generative AI - The basics
• What generative AI brings to software testing
• Using generative AI for testing - General principles
  • Practical Exercise #0: Getting familiar with the LLM workbench

2. Querying a large language model for testing - how to get good results

• Introduction to prompt engineering
• Prompting techniques
• Use cases with practical exercises:
  • Practical Exercise #1 - Designing test cases
  • Practical Exercise #2 - Improving existing test cases
  • Practical Exercise #3 - Test automation
  • Practical Exercise #4 - Analyzing anomaly reports
• Good prompting practices in software testing
• Summary and discussion of the know-how used

3. Managing the risks of Generative AI

• Hallucinations, Errors and Bias of AI
  • Practical Exercise #5 - Hallucinations / Errors - Examples from the testing domain
• Cybersecurity risks
• Environmental risks
  • Practical Exercise #6 - How much does it cost in energy?
• Other risks: Loss of skills, dependence on a third-party service
• Regulation of AI: The European AI Act

4. LLM-based applications for testing

• Applications for testing including generative AI
• Retrieval Augmented Generation (RAG) - Using company/project data
  • Practical Exercise #7 - Questioning a large document corpus
• Integrating generative AI into the testing process
  • Fine-Tuning of the LLM
  • Practical Exercise #8 - Integration of generative AI functions in a test tool
  • LLMOps
• AI agents for testing
  • Demo: Gérald - Virtual Manual Tester
• Summary and discussion

5. Summary: Putting into practice

• How to use Generative AI in practice
• Choosing the most appropriate AI model for testing
• Conclusion

Goals

• Master the basics of generative AI for software testing
• Correctly use the various techniques for querying an AI model for software testing activities (Prompt Engineering)
• Use AI to accelerate test analysis/design, test suite optimization, test automation and test maintenance activities
• Identify and assess the risks of using AI to accelerate software testing

Duration

2 days

LLM & AI-powered tools used in this training

• GPT 3.5 and GPT 4 (OpenAI)
• Mistral Small, Medium and Large (Mistral)
• LLM workbench, Application RAG, and Gravity (Smartesting)
• CodeLLama and LLama 3 (Meta)
• Sonar Small and Sonar Medium (Perplexity)
• Claude-3 Opus, Sonnet, and Haiku (Anthropic)

Target audience

• Quality Engineers, QA Managers
• Testers, Testing Consultants
• Developers, Test Automation Engineers
• Product Owner, Project Managers,
• Business Analysts

This course is organised and being held by our partner smartesting.