That's what mine looked like too (i.e. Name: message)
But the one difference is that I mapped everything to threads (i.e one doc per thread)
If you care about the entire channel and not just threads, then putting everything in one doc in sequential order makes sense
You can also customize the prompt to help the model understand what it's looking at:
https://gpt-index.readthedocs.io/en/latest/how_to/custom_prompts.html