Systems and methods for real-time, artificial intelligence control of an agricultural work vehicle and/or implement based on observed outcomes are provided. In particular, example aspects of the present subject matter are directed to systems and method that sense field conditions (also known as field finish) both before and after adjustable ground-engaging tools encounter the soil and that update a site-specific control model that provides control settings based on the observed anterior and posterior conditions. Thus, a control system can obtain sensor data descriptive of upcoming field conditions and can perform predictive adjustment and control of tools based the upcoming field conditions. The system can then use additional sensors to observe the outcome of the employed control settings. Based on a comparison of the observed outcome to a target outcome, the system can adjust for the next encounter of similar field conditions.

A method for making available a ballasting proposal for an agricultural tractor includes providing an implement for the tractor, calculating a target tractive power for the implement for each axle of the agricultural tractor, and deriving a target value for an axle load to be maintained on each axle based upon a function of the calculated target tractive power. The method further includes determining an actual axle load on each axle, comparing the target value to the actual axle load, calculating an axle ballasting value based on a result of the comparing step required to maintain the target valve, and outputting the ballasting value.