Effortless Game Testing With Planning Agents
summary
Planning agents built with symbolic (explainable) AI automate boring functional and regression playthroughs, claiming to cover 65% of testing and run thousands of agents concurrently with no onboarding. The system integrates with Unity and Unreal, records full repro steps and videos, diagnoses root causes, and supports golden-path, smoke and Monte Carlo runs. Benchmarks report about 17.5× per-dollar efficiency versus human testers; learned symbolic models scale in the cloud and have peer-reviewed recognition (KR conference award).
AI-written summary · about
chapters (12)
- 0:00AI-driven game testing intro
- 1:07QA shift to AI testing
- 3:39AI agents for regression testing
- 6:08Symbolic vs black-box AI
- 8:12Combinatorial explosion in games
- 10:22Early access testing limits
- 11:25Cost savings with AI agents
- 14:19Space engineering AI heritage
- 16:26Games best for AI testing
- 17:31Cloud scaling for symbolic AI
- 19:42Symbolic AI research advances
- 20:03Demo and case studies
related talks
18mSession: AI Effortless Game Testing With Planning Agents
05mArtificial Players in the Design Process: Developing an automated testing tool for game level ...
28mTesting at the Speed of Thought: QA in the Era of GenAI apps
31mIntroducing Agent Testing
08mReal AI Agents Need Planning, Not Just Prompting
30mAgents Building Agents
37mAI agents testing How to evaluate the unpredictable
46mAI Testing Beyond the Basics: Ensuring Truthful and Reliable Agents