Having some config trouble here… The Bot Performance view for the ‘Tier2_Support’ flow in EU-West BYOC indicates a 15% containment rate, yet the Conversation Detail view shows only 9% resolved interactions for the identical 2023-11-15 UTC window. The Architect flow utilizes standard NLP intents with no custom scripting. This variance impacts our SLA reporting accuracy. What is the definitive calculation logic for containment metrics in the Performance Dashboard compared to raw conversation logs?