Perhaps i should have explained better my reason for this. The way I envision it is Robots would pretty much be built solely for the manufacturing sector. When i think of "service industries" i think of things like: Grocery store, Clothing retailers, tech companies, customer support etc. All things that a 100% robotic population would not need (nor care about) for daily operations. Things they would need would include maintenance, energy production/resupply, software & hardware upgrades and support. All of which could be covered by "life supporting" industries.
If you think that should be larger then a minimum of 10% of the population i'll gladly listen to your thoughts.
There's plenty of manufacturing that is in what the game calls the "service sector." Things like trade goods are produced by populations, but take up no "manufacturing" workforce. So the sector split doesn't actually conform to the individual job function: The life support sector isn't just agriculture and life support; it's "all the ways in which things get harder to do as the planet becomes more inhospitable" - including any reduced efficiency of your mines caused by, say, the workforce having to spend time in a pressure chamber at the start and end of shift. Similarly, the "service sector" isn't groceries, retail, customer support. It's "all the non-military things you need more of (per capita) as the population grows." That includes things you physically need to facilitate expansion, such as rail lines, port dredging and so on. And it includes a lot of things your population needs to consume itself, or wants to divert to non-military uses. So there's going to be a lot of factories in that pool, they're just going to be producing for local consumption rather than the armaments supply chain.
The logic behind the "service sector" growing in the game (other than game balance) is that with a small population you can mobilize a much larger fraction of it for armaments production. This does not necessarily change just because you use a robot workforce; that depends entirely on your justification for this diseconomy of scale. But it seems reasonable to suppose that the effect arises from many sources, at least some of which would be unimpacted by the presence of robots.
Personally I model it as the smaller colonies outsourcing much of their internal consumption to production in the larger colonies, paid for by either direct subsidies from the government or the indirect subsidies from the wages paid to the workers in the armaments sector. As the colony grows larger, the cost/benefit balance of onshoring that production changes due to economies of scale, so it begins to do import substitution that takes up population to run. But your storytelling may have a different justification.
At the most extreme end, you could postulate that the robots do not have such diseconomies of scale at all, but you probably still want to retain the low-population "service sector" fraction. Just have it not grow with increasing population.
If you reduce what the game calls the "service sector," it should probably come with a corresponding malus to the production of Trade Goods, Wealth and anything else that population generates in Aurora by its simple existence. (Note that there is already sort of a model of this in the game, as populations below certain thresholds do not produce certain trade goods, though it is a quite rudimentary model.)
This idea actually opens up another interesting mechanic - the degree of society-wide wartime mobilization would impact the amount of workforce that could be diverted to the military-industrial complex, thereby changing the size of the "service sector," as civilian industrial capacity is mobilized for war materials production. Different populations might have different tolerances for wartime hardship, depending on how threatened they are feeling, how stable the regime is, what their general expectation of hardship is, and so on. Exceeding the level of military mobilization a population will tolerate would lead to dissent.
If you introduce that mechanic, it would simply mean that (non-sapient) general-purpose robots would be a perfectly pliable population, which could be set to any mobilization level. As technology introduces more sapient robot models (for reduced production maluses to research, and maybe some other things), it also increasingly constrains how much you can mobilize them to produce for wartime use rather than internal use before they go on strike. Basically, they become more like organic sapience.