How much are they really saving? Wouldn't it be a similar cost to just launch a new satellite than it would be to launch a larger servicing/refueling craft?
As someone mentioned if the launch vehicles can be reused then probably costs would be cheaper.
I think you also have to weigh the risks of launches. If a gov't or private company launches a replacement satellite and it blows up at launch or fails to reach orbit that could be very costly. If the satellite in orbit is working fine, then just launching some fuel or spare parts could be much less risky financially than launching a full replacement satellite. Along those lines, reducing the potential financial risk of launches can reduce launch costs throughout the industry. If a launch fails someone does pay for it somehow (e.g. a private reimbursement/guarantee, insurance policy, gov't backing, etc.) and one way or another that cost will be reflected in the larger costs of the satellite launching industry, so reducing financial risk of some launches should reduce costs everywhere.
Keep in mind too the article states this robot could potentially be used to gather or redirect space junk or asteroids, which despite sounding like long-shots could be very useful, and even an incremental step towards those goals could be valuable.