John Calvert, the country's principal exponent of ID, answered that question in a treatise he presented to the Ohio board. Calvert described the "methods" by which scientists can "detect" design in nature.
In summary, if a highly improbable pattern of events or object exhibits purpose, structure or function and can not be reasonably and rationally explained by the operation of the laws of physics and chemistry or some other regularity or law, then it is reasonable to infer that the pattern was designed. — the product of a mind.
That, in a nutshell, is ID. It offers no predictions, scope modifiers, or experimental methods of its own. It's a default answer, a shrug, consisting entirely of problems in Darwinism. Those problems should be taught in school, but there's no reason to call them intelligent design. Intelligent design, as defined by its advocates, means nothing. This is the way creationism ends. Not with a bang, but with a whimper.