It would take me and nicco probably 3-4 months mastering the Polygon Runway level of detail and Character rigging. We ventured into this and came up short. We will be using Evee (a less realistic but low CPU friendly renderer) and following the Blender Non-Photo-Realistic (NPR) principles rendering. 
We wont be starting characters until maybe after the ships, so probably 2024 (if we are around that long and we continue to grow. Small Businesses Fail 80% of the time after 3 years and anything can happen to Nicco and Myself that would cripple us we are in our 40s).  

GMs will be using the models and the HOST of characters we will generate, since Blender’s Procedural Generation tool will eventually allow for us to automate a lot of character modeling. Basically the GM can direct artists or players to get our assets for posing. 
Basically – Blender characters in low poly and cell shaded for character scenes and portraits. Nicco doing commissions but more likely – GMs finding many artists to use the 3d Assets to make them more customized scenes for the game. 
Blender being so powerful, it is going to be no surprise that Facial Capture Avatars would grab these low res models and animate them for a game. Like the Planet Generator – the GM just wants a random set of NPCs to generate and screen shot or render for the game.  

AI + Procedural tools in blender would mean eventually someone will make a plug in that will allow a NON-blender user to pop out a character quickly. 

As AI and Tools become more powerful the way GAMES will develop is that GMs and Gamers will have more visualization tools in 2025-2030. In around 2029 people will be having Avatars ready for the game and the GM procedurally generating a Cast of NPCs from Blender. Cities, Worlds, and Everythign are Procedurally generated. 

Even the GM doesn’t know everything going on and is in for the Adventure. By 2030 the new generation of tech savvy TRPGers will have huge WIkis of cast of characters and worlds. Using AI to populate the Imagery – and the next frontier for AI is generating procedural Rigged Models in an Open Source Low Resources way. Someone with an I3 or Ryzen 3 spends 30-90 minutes prepping for a game and generating 1-2GB of content to show the players. 

I can be wrong it will be Earlier or Later. The thing is Mneme and Game in the Brain will be one of the main sources of assets and targeted investing (Investing the money from the Books to make more of this content). I personally help people set up their Blender and their products for games or YouTubers just making tutorials how to set it all up. And Nicco and I borrowing from them to make the products even more user-friendly. 

Probably a few hundred gamers would buy our stuff yearly to support us, and will get models which they can render and play with. That get them thinking “I’ll make a planet and space station here and a Ship bursting out of the dock and the PCs are going to have this scene or this was what happened a while ago.” 

If I had more resources I would do something more ambitious like put this all in a GODOT engine but that’s assumes a thousand people or 10,000 people support us in our price of 30-50usd a year because of products we make.  for that to happen someone online must have tried our stuff, which I’m sure is flawed, given feedback and we polished it up and people will affirm it when they check it out (Since most of the product is available for inspection before purchase). 

There will be many people making regular RPGs – and Our products will complement them! of the 1000 rpgs there are with the sandbox nature of Cepheus Engine – there will be people taking the stuff here and using it in their games.