{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/b6fb5cd87e5a458385541d9e994bba04\" frameborder=\"0\" width=\"1920\" height=\"1440\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1440,"width":1920,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1440,"thumbnail_width":1920,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/b6fb5cd87e5a458385541d9e994bba04-7d0a9014e471afae.gif","duration":118.869,"title":"AI Reconstructs Evidence From Depositions","description":"This Loom demonstrates using Gemini Fog and Argon to extract evidence from a deposition and reconstruct events accurately. The author describes analyzing a 212-page, nearly three-hour deposition video in one call, pulling out vehicles, timing, and where each witness stated they stood, then building a dynamic 3D world as code and checking it against the evidence to fix its mistakes. They review findings showing a witness claimed her view was unobstructed, but video analysis identified the exact moment when a bench shelter and a turning truck blocked the call completely using a plane geometry check reproduced by 23 agents. The workflow is extended with a new witness to confirm that new information feeds back into the model, and the lawyer can export a clean courtroom-ready report."}