LLMs don't generate that sort of templated response for conversation (although they do for software code).

It's more likely a traditional ad, using some classical algorithms perhaps.