This was deep research, not the regular chatbot. In theory, it should have read all the documents, but you never know