Airflow's 3.1 release introduced a groundbreaking feature—human-in-the-loop interaction—allowing users to incorporate human input into workflows, expanding use cases in AI and operational management. The Astronomer DevRel team tested this feature using the AI-assisted Astro IDE, which successfully implemented several new human-in-the-loop operators, demonstrating the IDE's understanding of the update despite minor errors that were quickly rectified. The Astro IDE's performance was notably superior to other AI agents like ChatGPT and Claude-4.5-sonnet, which struggled to correctly implement the feature. The test highlighted the importance of context and documentation in AI's utility, with the Astro IDE benefiting from updated guides and examples. Further improvements are anticipated as new information is integrated into the system, ensuring the IDE remains a valuable tool for developing Airflow deployments.