from what I understand, the main problem with a "conventional" satellite is that the transmitter beam is fixed and the shape of the antenna physically determines the shape of the coverage, which cannot change once the satellite is in orbit. if you use a generic shape for the antenna dish and multiple beams,then you can control the shape of the coverage digitally by changing the amount of each beam.
Its touched on briefly in the article but this is what phased array antennas are good for. With a sufficient number of elements, you can modulate the shape and direction of the beam electronically. Its not magic, so you're still operating within some envelope of transmit power and receive sensitivity, and the antenna's effectiveness is going vary across the spectrum, but you do get a few new knobs to twist.