In the presented process of preventive planning of Product-as-a-Service offers, a solution competitive to the classic resource allocation model was implemented. Unlike the conventional formulation, which searches for variants of allocation of sets of devices of various types, with parameters that meet the expectations of the customers ordering them, a genetic algorithm (GA) implemented in the proposed model looks for the structures of the so-called stepping crawl threads (SCT) reconstructing sets of ordered equipment in subsequent steps. SCTs initiated at the starting point of the Cartesian product space of the functionalities repertoire of the offered equipment penetrate it in search of offer variants that meet the constraints imposed by the budget sizes and risk levels of individual customers. The conducted computational experiments, confirming the evolutionary nature of the proposed model, indicate its competitiveness compared to conventionally used ones, allowing for a 10-fold increase in scalability.