It would really help to create better knowledge data, if we could see what part of knowledge (what text) the chatbot recieves. It would be even better if we could configure it, because for eg. e-commerce store, if I make a text file from each product, with url, price and description, it just cannot cut it in the middle. Full file context needs to be passed to the LLM as a context. It won't be long anyway. Also, option how many items found in the knowledge base should be passed to the ai model

Give the ability to see what you put in the knowledge center. I need to make sure the information in the bot is correct from time to time. A PDF or document is up loaded but I am not able to open that document to determine if it is the most current information. Provide the user the ability to open that document to determine it's content is current and usable or if it needs to be replaced.

Make possible, that we can track what we have provided for training. When we upload documents or websites, it does not show what we uploaded. This makes it very difficult to update things later or to track what and which versions we send for training. The files should stay there and be accessible or downloadable. If not possible. at least the name of the website url or the document name should be there,.

Adding multiple URLs at once

Renaming knowledge source (especially free text)

Ability to view and edit uploaded knowledge sources.