Don't look for customers for your AI / NLP technology, look for customers with problems that can be solved with your AI / NLP technology and then sell them a solution to that problem.
That's probably your best bet for areas to start investigating. Seriously. Smart people tend to think they have to solve hard problems, but usually (in my experience, and from watching other successful entrepreneurs like Patio11), the money's in doing stuff that you think is really easy, and other people think is hard.