Stream Assistant Response

I would be interested to understand how have you replaced the AI Assistant ? Have you gone to chat completion? I’m not sure how llama-index has helped you, perhaps your use case is very different than mine