{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/86a98fc841224058948d1dfeafaa718a\" 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/86a98fc841224058948d1dfeafaa718a-32a449e171e04c9b-full.jpg","duration":3712.88,"title":"QA for High-Stakes Financial Apps: Insights from Phantom. Hosted by Mobot.io","description":"This Loom discusses how Phantom scales and modernizes its QA and release testing, including the role of AI and real-device automation. Jake and Kevin describe their backgrounds and explain how Phantom runs a tiered pipeline: local and device simulator builds, PR unit and component tests, smoke tests on real devices, nightly full regressions (including MOBA-ledger coverage), plus production synthetic and load testing, with weekly releases on mobile and CD for web and backend. They cover how hot fixes differ by targeting the minimum code delta and running a focused subset of tests, because app and extension rollbacks are difficult due to store review cycles. They emphasize challenges from AI-generated code that increase PR review and testing load, improved triage with AI reducing issue investigation from hours to about 15 minutes, and the need for test strategies to adapt continually."}