For reverbs, I use Seventh Heaven, a fantastic Fusion-IR emulation of the Bricasti M7. So it's convolution based, but lets you manage reflections, decay time, and other neat parameters. My template is set up like this:
1. All tracks by default are set to output to a Null bus (-inf dB, so silent).
2. Five send/busses as follows.
- Dry - This is a 0 dB unity gain bus by default with nothing on it.
- ER - This has a subtle low roll-off into a reverb configuration set only to early reflections.
- LR - This has a less subtle low roll-off into the identical reverb configuration, but set instead to late reflections.
- Amb - This is like the LR send, but it has some more creative effects like Valhalla Shimmer adding a spacey, lasting shine. I don't write strictly traditional acoustic orchestration so this is a personal thing.
- Sub - This is like the Dry send, but it has a low pass at around 100 Hz. This send is to artificially increase and manage the low end of my instrument tracks. In this send I could do things like stereo field management, compression, automation, etc.
3. All tracks in my DAW have the send faders available so I can control the blend on every element in my mix, right in the DAW mixer without opening plugins. I can place a choir further back at say, 10% Dry - 30% ER - 60% LR. For soloists, give them Dry detail and some Amb to add space without pushing them back into the room. It's case by case; this template is about allowing me easy access to use my ears and adjust things, not necessarily about pre-mix ideologies.
3 (sub). I have a few reasons to use a Dry send and not simply output the track to Master. One is that it allows me independent control of the track fader (which controls all of these sends together, because they're post-fader sends) vs. controlling the level of dry signal for blending. In other words, without doing this, I'd have to use the track fader to control Dry signal, making it useless for general mix adjustments and automation, and then changing the other sends to pre-fader so I can pump more signal into them should the Dry have to be really low (for spacier/further sounds).
Another reason is now I can process the detailed, clear parts of my mix without also processing the reverb. There are relatively few cases where I actually do this, but it's useful in some circumstances. One time, I took the Dry signal and used it as a key input to a sidechain, so that I was driving the master compressor only by detailed information and none of the buildup from room sound. I like to experiment with unconventional mixing techniques, sometimes it pays off.
EDIT: Worth mentioning, the Sub bus is completely dry, so if I shove something in the back of the room, like the string section, I can still steal their dry low end. It's one of those "larger than life" approaches to mixing.
4. All sample libraries have unloaded all mic positions except for the close positions. Exceptions made occasionally for drum overheads/rooms or certain libraries where I like the smoother sound of a slightly farther position (like Tree mics in Spitfire libraries). This isn't a hard fast rule but it's an important starting point for the most efficient RAM usage and easiest mixing process. Having all close mics and managing reverb through just a small number of plugins is not only incredibly efficient, it sounds way better than anything I used to do before, and blending libraries from different developers is a completely seamless thing.
Here is an example of a song I mixed utilizing this workflow. There are sample libraries from 4-5 different developers here, yet it's not even a thing to consider when mixing using this approach. There's barely any EQ happening because they're quality libs, just gentle filtering to control ranges.
Excuse the scratch composition; additionally, this demo was made before I got Seventh Heaven. It's the TSAR-1 from Softube. It would probably sound even better if I replaced the reverb config.
Here is an "alternate mix", where I arbitrarily changed the positioning of elements. I brought the choirs closer and moved the strings back for a more intimate sound, and this was done purely through managing the Dry, ER, and LR faders on those elements (celeste track, string bus, choir bus).
Here's a snippet of the Menuet from Ravel's Le Tombeau de Couperin. I just snagged the MIDI online from somewhere, so the sample sequencing probably isn't the greatest. But this is demonstrating combining Spitfire Strings and Berlin Woodwinds, and it's totally seamless.
EDIT: Make sure you know if your sends are pre or post-pan. Mine are pre-pan, which renders the pan control on mixer tracks unfortunately useless. I simply instead use the pans on the send faders or have a pan effect in the FX chain of the element. If you can make your sends post-pan, that's even better.