I find that artificial intelligence, when it finds information in the database, doesn't give a very "human" answer. It often answers with the text as written in the database. Is there a way of making the response more "human"?🤖

2 options. You can try and use the personality agent and prompt it to write as a human would. You could also rewrite the data in the KB to make it sound more human ish but there will still be an LLM that sees that text and writes an answer so the best option is the first one

Thanks. I was thinking of the first option but I can't see where to configure it.

Thanks, I tested it but I don't think it changes the text, I put in personality "always makes short sentences of maximum 30 words" But when the AI finds the answer in the database, it gives the complete answer point by point. But if I've understood correctly, when the personality is activated, it's for the whole chatbot? And what does the "Enable Personality Rewriting" option mean? Thanks for your help

It will rewrite the outgoing message to the user. Try telling it that it's a funny clown that always adds a joke at the end of its sentences. Enable personality rewriting just enables it to rewrite messages. You can disable the Personality agent on each node and change its prompt

Thanks, I'll try that. Can you show me a screenshot of the option to disable the bot on a node?