Do you have a standard suite of conversation topics/messages that you A/B test against prompts/models?
Do you have a standard suite of conversation topics/messages that you A/B test against prompts/models?