Gjore Zaharchev: AI in Software Testing Expectation vs Reality
summary
AI-generated test cases often fail until prompts include exact constraints: replacing a generic request with a prompt with exact field constraints (email username 20–100 chars, password 14–40 chars) turns useless combos into usable tables. A REST API example grows coverage from 3 to 28 tests, catching bugs missed before. Tools that code scan for flaky tests flag whether failures are test or product issues, integrate with Selenium/Playwright/Cypress, and require technical prompts and data-handling rules to avoid leaking business data.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 3:14AI testing expectations vs reality
- 7:53Fear of AI misconceptions
- 11:05Practical AI tool use
- 15:23AI-driven test design
- 18:12Intelligent defect detection
- 21:52Focus on application under test
- 25:09Skills for AI leverage
- 29:24Data leaks and confidentiality
- 31:09Empowering teams with AI
related talks
44mHow AI is Exposing the Existential Crisis of Software Testing
29mQA vs AI: Friends or Foes?
32mQuality engineering in the AI age: From testing AI to leveraging GenAI
01h 05mPanel Discussion - The Future of Testing: Impact of AI in Quality Assurance and Beyond
40mFrom Bugs to Brilliance: Leveraging AI in Modern Software Testing
42mThe Power of Human and AI Collaboration in Software Testing
35mApp Testing with AI Tools – Make it Easy
14mPrompt Engineering Impacts to Software Testers : Staying in Demand