Building a custom widget that needs live conversation details. I’m torn between using /api/v2/conversations and /api/v2/analytics/conversations. The v2 endpoint feels right for real-time state, but the docs for analytics are vague on latency. Does the analytics endpoint actually cache data, or is it just a different view? I need sub-second updates for screen pops. Tried hitting both with a GET request. The analytics one returns 200 but the data feels stale compared to the main conversation API. Which one should I hit for live agent data?