Hi all
I’m attempting both the openai conversation agent and the extended conversation agent but both are letting me down.
For the standard conversation agent integration, I continually get a failure on intent recognition unless I ask it the most basic thing (e.g. What is the time). I’ve had issues with this integration linking in to openai before so it could be that. So I tried using the extended openai conversation agent integration and it was working fine (using same api credentials as the other so go figure). But recently it’s taking my commands literally. For example, if I ask it to turn the bedroom light down, I will get a response saying it couldn’t find the device lights.bedroom.light. It used to find what light devices I had in that room and use that, not take my ask word for word. I can’t remember all my device names 
Im sure I’ve probably mucked something up. Could anyone let me know how I can look to resolve this?
Thanks