UCAVs have worked pretty well for CAS in low-intensity, counter insurgency type conflicts but they are totally dependent on satellites for control and navigation. Potential near-peer adversaries are investing heavily in as anti-satellite weapons so in any future high end conflict those satellites will probably be the first casualties.

The US military and allies have been working to develop "loyal wingman" type UCAVs that can operate under the control of a crewed aircraft. Those show promise, but still require the controller to get fairly close in order to allow for reliable communications in a hostile EW environment.

Autonomous flight control software is nowhere near being able to perform CAS. That's probably decades away.