A session by Shrinath Rao explaining how to build an autonomous QA system using a “Council of Subagents.” It demonstrates how AI-driven roles (like orchestrators, analysts, and engineers) collaborate to automate software testing, improve coverage, and keep pace with rapid development.
Understand how an autonomous QA system operates using AI agents
Learn the concept of a “Council of Subagents” for distributed testing
Identify the roles of orchestrator, analyst, architect, and other agents
See how to break down features into structured test plans
Learn how automated test execution is managed by AI agents
Understand how failures are detected and auto-healed
Explore how end-to-end QA workflows can be fully automated
Learn how to scale test coverage with minimal manual effort
See a practical implementation using Claude-based agents
Understand how AI can accelerate QA in fast development cycles
In this session, Shrinath Rao presents a structured approach to transforming traditional QA into an autonomous, AI-powered system. Using Claude-based agents, he introduces a “Council of Subagents”—a coordinated network of specialized roles including orchestrators, analysts, architects, engineers, sentinels, healers, and scribes.
The talk walks through a six-phase pipeline, starting from feature analysis and test planning to execution, error detection, and automated healing. It highlights how each agent contributes to a specific part of the QA lifecycle, enabling scalability and reducing manual effort.
A live demo showcases how the orchestrator sets up workflows and how subagents collaborate in real-time to generate, execute, and refine test cases. The approach significantly increases test coverage and helps QA teams adapt to fast-paced code changes.
The session concludes with practical insights, manual testing integration, and next steps for implementing such systems in real-world environments
