The scenario setup can take a number of different 'cases' with differing strengths and starting locations (while still being essentially 'the same') the reinforcements will also arrive in their own sweet time.
Once the forces are in play (and sometimes there *is* only one setup, or no reinforcement), the AI has the starting position, and a set of 'operational objectives' - terrain features to take/hold/occupy/deny/withdraw to, and the destruction of enemy forces. While this means that any given situation will play out 'according to a structure' the individual movements to accomplish the various tasks, and the strength of AI resistance in each sector will vary significantly from one run to the next, with the AI actually being quite competent and aggressive (but within sensible bounds most of the time) the same AI acts as a layer within your own forces, where the player essentially gives the grand-tactical orders between the 'operational objectives' and tactical actions on the ground that accomplish these tasks.
The tutorial is perhaps one of the least varying scenarios, and yet it seldom plays out the same way even over one day, and is 5 days long. There is a certain 'arc' of action/reaction and course of action that is expected (A US Arm'd Division hitting a single VGD Regt (+) of 3Bn doesn't leave much scope for the dramatic and unexpected swing - but still the situation on the ground when the German mechanised forces arrive, and what they can accomplish to slow/stop the Divisional attack is quite varied, depending on whether the advance mission is complete or still in progress, and how much of the force can be tied up by the German Defence (note this is usually played as Allied).
The 'course of action' does tend to have a similar flow if similar decisions are made, but neither the AI nor player are likely to make identical actions and the end state varies a considerable amount.
Most of the time the attacker has sufficient forces to 'guarantee' his advance, so that is seldom in issue, but how fast and how costly this advance is does vary, and the force preponderance can alter over one scenario and the decision to halt the advance and switch to a defensive posture is often needed (or vice versa). Also to keep your artillery protected and supplied (Of course you want to do this for all forces, but artillery is the God of War and frequently the bulk of the combat ability of a formation, and needs much more supply to keep it active...)
Local counter-attacks are often required (and made).
Probably try the demo, and see if you can consistently destroy the 12VGD(+) forces ~ this is possible. See if you are happy with the different approaches to delay the AI attempts, and the options to control your forces. Note that this is very unbalanced as a scenario, and the AI deliberately isn't given much to work with to begin. It should be played with realistic or painfully realistic orders delay and standard reinforcement and supply options at first ~ you can add an additional Bn of VG infantry and improve supply delivery if you find it a bit too easy.