You should make whole case with silicone.
There is an idea that I have been meaning to try where instead of having a keyboard "sandwich" be suspended on stand-offs or via a plate's edges, each switch's central "centre pin" would transfer its energy into a bump in a silicone "core" directly. I'm curious to what the sound profile of such a keyboard would be.
I've been thinking that perhaps for a very small keyboard with low-profile switches (think split 40% ergo with Kailh Choc), the PCB might provide enough rigidity alone so that the silicone core would be all that would be needed as both damper, feet and low case walls. This could also be a cost-effective way to make it a tented case.
One issue with this idea though is that a flexible case would not protect the components against blunt force — which a rigid case would. So perhaps some rigid part might need to be part of it anyway. Those horrible rolled-up silicone keyboards tend to have a box around the controller for this purpose.
I also think that maybe it would need to be co-moulded (double-shot) with two types of silicone: one shot with a harder type for feet and case sides, and another shot with a softer silicone for internal damping. Perhaps rigid parts could be laid into the casting in-between shots of silicone.
Both types of silicone would need to flow well, which means that you'd need two types of silicone: not just the same type with/without "thixotropic" additive — because that would make it thicker and more difficult to pour.