video1.0<iframe src="https://www.loom.com/embed/dba6b4b722a84944b26d8c48b4530014" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/dba6b4b722a84944b26d8c48b4530014-b9c5dec25f39ce69.gif1592.758This Loom demonstrates Prod.IQ, an AI production planning agent for manufacturing companies that simulates scenarios and generates data-driven plans. The creator shows a Docker-based setup with a React TypeScript frontend, FastAPI backend, PostgreSQL, automated CSV column mapping using fuzzy matching and inference, bottleneck detection, weekly action timelines, narrative manager summaries, and downloadable six-page PDF reports. For performance measurement, the agent uses a five-layer scoring system including data specificity, bottleneck accuracy, and action specificity, and is benchmarked above 8,000 versus a plain Grok score around 2,700, with a layer breakdown of 1,500, 2,000, 1,581, 1,000, and 2,000 marks. The video also covers security practices with no API keys or secrets committed, loaded via a .env file and excluded from the repository.