Where does your company draw the line on sensitive data in AI tools?