Creating course schedules at a university is a complex combinatorial optimization problem in which teaching activities must be assigned to available time slots and resources while satisfying organizational, spatial, and instructional constraints. This study presents an empirical analysis of a periodically conducted department-level timetabling process supported by an integrated environment developed by the authors. The environment covers the full planning cycle: automatic acquisition and standardization of data from dean’s office systems, preparation of input data and constraints, timetable generation using dedicated software, publication of final schedules, and analysis of process data. The empirical material covers eight consecutive semesters and thirty repeated generation runs for each semester instance. Normalized indicators were used to compare instances of different scale, including constraint density, generation time per constraint, generation efficiency, and post-generation adjustment effort. Across the analyzed period, 39,977 constraint records were processed, of which 75.74% were time constraints. Manual adjustments decreased from 154 to 52, while the adjustment rate fell from 9.90% to 3.04%. The results indicate that instance size alone does not explain the practical difficulty of timetable generation. Constraint structure, generation-time stability, and post-generation correction effort provide complementary information for assessing academic timetabling complexity under real organizational conditions. The observed reduction in manual adjustments is consistent with improved process control, although the available data do not allow the observed trend to be attributed solely to the integrated environment.