How ANZ Leverages GenAI to Transform Testing Activities
summary
ANZ describes productionizing GenAI to ingest internal docs and generate test requirements, scripts and realistic data, running models behind the firewall on an inso platform orchestrated by a LangChain agent and mixing Gemini Pro with a fine‑tuned LLaMA and other models. API testing shows a measured 72% speedup, environments self‑heal and surface faults, and payments run ~800 tx/sec so AI focuses upstream on requirement verification, risk prediction and test scope reduction rather than final safety checks.
AI-written summary · about
chapters (21)
- 0:00GenAI in ANZ testing
- 1:12Why ANZ embraces AI testing
- 8:34Early AI testing experiments
- 10:59Retail migration AI opportunities
- 13:47Payments AI test automation
- 17:18AI for testable requirements
- 20:32ANZ’s AI models and tools
- 24:08Scaling AI in testing teams
- 27:57AI for migration testing
- 31:53Trust and limits of AI testing
- 32:55AI for realistic test data
- 35:01Future of testing jobs
- 37:48Team changes with AI
- 40:37Resistance to AI testing
- 42:03Measuring AI testing gains
- 45:50AI in payments testing
- 48:17Accountability in AI testing
- 51:02AI leapfrogging manual testing
- 53:30AI hype and hallucination risks
- 55:19Regulatory challenges with AI
- 59:28Q&A
related talks
29mIndustry generative AI in action
28mUnlocking the power of GenAI
23mScaling GenAI Applications and Agentic AI for Business Competitive Differentiation
41mAI-Augmented Testing: How Generative AI, Prompt Eng Turn Testers Into Superheroes
17mAI-Powered Integration Testing
24mAIGenOps: Integrating Generative AI into Platform Engineering
29mAccelerating GenAI Innovation: Lessons From Intuit's Agents an...
32mQuality engineering in the AI age: From testing AI to leveraging GenAI