Comment by smartician

If the goal is to emulate human behavior, I'd say there is a case to be made to build for the same interface, and not rely on separate APIs that may or may not reflect the same information as a user sees.