You might be interested in Stellar Monarch (https://store.steampowered.com/app/446000/Stellar_Monarch/). I haven't played it (no linux version) so can't recommend it but it seems much more focused on role playing as emperor sitting in your throne room. I think a successful mobile CK2 would look more like this and less like the desktop map staring one.
1. Each tier is its own game, 2. The games are not independent
As a simple, two-game, example. Imagine a strategy game such as Shogun (turn-based strategy). When the players kn that game start conflicts that opens up a match for FPS-type players to join. The outcome of the battle will depend on the FPS-player performance while the match-up is skewed depending on the relative forces deployed by the generals.
Maybe cities are maintained by more Minecraft-esque players, while the eye in the sky can submit tasks down below for resources, certain buildings, etcetera.
What would be in it for the player extras on the tactical level? It's hard enough for level design and matchmaking to compose a match of random internet players when there are no further constraints, forcing the parameters of one players strategy session on top of that problem can only make it worse.
"RPG sim" would aim for the opposite, creating a uniform UX with consistent interaction pacing (even if on wildly different timescales) for all the off-screen sim layers.
For example a pirate game would have a detailed social sim of the crew, an "FTL level" sim of the ship in combat, a "Sid M. Pirates" level sim of ship to ship combat (that is all NPC off-screen when the player character isn't in command of a ship) and a sim of baroque-era Caribbean economy. A low level sailor would just manage relationships and balance focus between categories like following orders, initiative, camaraderie, bravado and risk avoidance (each influencing how he fares in the off-screen social and "FTL combat" simulations), peppered with the occasional decision set-piece, whereas a pirate-king might decide on resource allocation and strategic city upgrades (but still not micro-manage anything but relations). The tactical level would be a much more interactive (if the PC happens to be in command) series of decisions, but still do exclusively fast-forward-to-pause timing. ("free form turns", imho mobile games should be as passive as a book whenever you put down the phone)
The pattern could be applied to anything that has a somewhat limited set of characters for the social layer (the CK2 courtier population might still be a bit heavy for mobile...) and involves separable tactical and strategic layers. The original idea came from the question of how one could do a WW2 air combat sim on mobile that is deeper than Sid M. Ace Patrol but unlike IL-2 still without a dexterity factor. The player would balance focus between e.g. spotting, engine operation and navigation when on mission but not in combat and during engagement between aiming, evading, energy conservation, situational awareness and so on (plus a series of decision points that might be seconds or minutes apart depending on situation). The social layer might involve storytelling decisions like choosing to "accidentally" miss an emergency scramble after a night of heavy stress relief or flying under the influence.
With a bottomless fantasy budget, this kind of game would come with full 3D modeling for illustrating every sim layer, but only ever show a single frame at each interaction, representing the state of the simulation (which is actually closer to board-game abstraction than to 3D physics) in a semi-randomized way. Reasoning: animations quickly get old no matter how good they are and single frames can be doused in nice-looking postprocessing that would cause unsightly artifacts if animated. An endless pool of informative flavor illustrations, not the usual showing-off of GPU power and production effort.