When glaciers (in general) are melting enough to be a significant source of sea level change, their flow (in general) is not going to be sustainable. If you irrigate a desert with a water source that is running out, you may make things better temporarily, but it is not likely to be the best solution.

This only holds true for glaciers as a class, of course. You may be able to identify some glaciers that have significant runoff but are not shrinking, and you might use these productively, and in doing so help make up for some of the sea level change caused by those glaciers that are melting themselves out of existence.

More specific to your question, I can say that the glaciers in the arctic and Greenland are not going to be useful in watering deserts that are conductive to significant plant growth (too cold), and many of the other glacier networks, such as those in the Himalayas, are already overused -- for example, the Gangotri glacier network, which is one of the major sources of the Ganges, is fairly well maximized in terms of basic flow, to the point where there is a very complicated treaty in place to make sure that India doesn't gobble up too much of the water before it reaches Bangladesh. There are still floods in monsoon season, though, so that would be the area in which to work on water catchment and redirection plans.

As you might expect, the situation is different in each glacial system, but I don't know of any offhand that are underutilized and near a desert. I suspect that this is the case specifically because people in deserts want water, and redirected water to their benefit long ago.