Case studies

OSCAL.AI manages large vehicle fleets serving thousands of collection points across multiple service lines, including used oils, cardboard, paper, rubber, and specialty liquids. Faced with increasing margin pressure and stricter environmental requirements, the company needed to evolve its platform to maintain operational performance and reinforce its position as a technology leader.

Several operational limitations were restricting the platform’s growth potential. First, collection routes were generated through a simulation-based approach that tested different route combinations, rather than using a true mathematical optimization engine. Second, customer prioritization relied heavily on manual decision-making, which reduced overall efficiency. Finally, multi-day fleet planning remained highly complex, leading to underutilized trucks and collections scheduled prematurely relative to customers’ actual inventory levels.

As a result, nearly 58% of collections were performed too early, while only 34.5% occurred within the optimal profitability window. This lack of precision increased the cost per kilometer, added unnecessary carbon emissions, and limited the overall value the platform could deliver to its customers.

To address these challenges, OSCAL.AI partnered with Vooban to develop a custom route optimization engine.

The solution is built around several core components. The first is a prioritization model designed to manage three types of collection tasks: dynamic collections automatically triggered based on inventory levels, recurring contract-based visits, and fixed-schedule pickups. This component also incorporates detailed customer inventory monitoring, including daily accumulation rates, target inventory thresholds, and overflow prevention limits. Together, these capabilities enable proactive demand anticipation and more effective prioritization of collection activities.

The second component is a multi-compartment optimization model. This capability allows a single vehicle to transport multiple incompatible products, such as used oil and de-icing fluid, within separate compartments. The optimization engine dynamically assigns products to the appropriate compartments while respecting capacity constraints and maintaining balanced vehicle loads.