Every now and then a customer asks if it’s possible to have a tool that flags dodgy responses based on gibberish or bot responses and so on. It would be good to build something that can at least help to avoid the user having to go through each response individually.