There'll be numerous factors and considerations here. Bus to send several things to vs as an insert or in the pan slot, solo vs ensemble... are you using the convolution or just the late reflections... etc

I'm influenced by the way the original Hybrid Reverb used diffusion, 0% is as tight as possible, 100% is as diffuse as possible; and I automated that. So an instrument with little spread, from the power panning is reflecting a certain way, more locally.

And there are compositional approaches, eg., the last sustained thing is more diffuse than it was during the buildup, similar to more wet level. I use it individually with individual settings quite a lot of the time; or if it's in a bus getting sends I'm using only the late reflections sweetening the tail. Or like mixing Teldex semi-baked into that Berlin product with Synchron, the Teldex-colored channels get some Synchron tails, while the enveloping mixer channels are output separately and dealt with judiciously. That's gone, Synchron enhance, & I cannot fathom a point to getting rid of it