Client : AXA IM
For some questions where there is relevant content available in the library but the questions is framed in a different way, the AI modifies the response. Can we give instructions to the AI to use the exact response from the library without it modifying.
