We need a way to store Safe Words that AI can use "as is" rather than being blocked or masked. The example is that we recently had a fictious user demo story and we used a popular sporting goods company name. AI blocked the name always showing Xxxx's Sporting goods. We would need a way to configure all true allowed names that when used in context with the proper phrasing and even pucuation can be allowed even if, when taken out of context, might imply that AI should mask the letters. We need some way to configure these safe words with a customers approval and acceptance that the name could be misconstrued and the customer takes full responsibility.