AI agents are already designing, coding, and testing embedded device UIs today. How can we enable them to also test UI interactions during runtime?

AI agents, such as Claude Code, Qwen Code, or GitHub Copilot, can already create unit test cases, trigger visual UI checks, and compare running UIs compared to UI designs. Embedding a minimalistic MCP tool to a Qt application allows AI agents to take real world testing to the next level. This presentation outlines what a locally-running Qt Interaction MCP tool can do and how it opens the door to multi-modal experiences in the future.

Speakers:

• Peter Schneider, Principal of Product Management, Qt Group