AI-Generated API End-to-End Tests: A Legacy Rewrite Case Study
summary
Arolla engineers present a case study of using AI to generate end-to-end API tests to secure a legacy API refactor handling €500M/year. They describe an AI workflow built from agents with a multi-cycle planning (MCP) approach to reliably regenerate large JSON comparison tests (>200KB), demonstrate prompt-design producing prompts that generate code, and explain how strict code-quality rules and context-separated sub-tasks reduce long-term AI costs.
AI-written summary · about
chapters (10)
- 0:00Legacy challenge
- 6:03Manual test limits
- 11:01Legacy API challenges
- 17:58First AI test approach
- 22:34AI code generation
- 26:48Team adoption
- 31:22Deterministic generation
- 34:10Jazz analogy and PSE
- 37:19Reevaluating test practices
- 39:44Conclusion
related talks
32mQuality engineering in the AI age: From testing AI to leveraging GenAI
25mRethinking Software Quality: Why API Testing Needs to Change
24mThe Power of Generative AI in API Product Testing and Validations.
20mPower of OpenAI in API Testing
47mNext generation of API testing : Unleashing the power of AI
37mAgentic AI and the Next Decade of Quality Engineering
44mHow AI is Exposing the Existential Crisis of Software Testing
26mAI and APIs: Ensuring Platform Migration Reliability with AI