Remember that your goal is to create a series of simulations representing a whole real-life problem. The process is much easier when the whole real-life problem is small. I mean that the problem could demand from a person a little time and few skills to solve the whole problem from the start till the end.
For example, driving a car could take 10 minutes; preparing the presentation and pitching it could take several hours. It means you can give students a simulation of this real-life problem in the first lesson! Yes, it will be in an elementary condition, but still, the student can make the first whole real 10-minute drive like a real driver (in a very safe and easy condition); or one can prepare and pitch the whole real presentation on the first lesson again in an elementary condition.
And what if you want to prepare an architecture?
The month could pass from the start of the architecture problem, e.g., designing and building a bridge, till s/he finishes it! If we ask students to design and develop the whole building by simulating the real steps through which architecture is going (even in the most manageable condition), it will take too much time for the first lesson. This first lesson could last for half a year or even more! Imagine what a stressful timetable will be for the students: "The first lesson starts on 1st September 2023 at 10 a.m. and finishes at 12:00, 1st September 2024…." Of course, there is no chance for repetition of such a simulation so that the student could go through it several times per program.