In the attended home delivery industry, customers order products online, which are then delivered to their home within a time slot of their choice. Given the nature of the products, e.g., groceries or appliances, the customer must be at home to receive the goods. Therefore, on-time delivery is of key importance. Due to the inherent variability of various planning factors, such as uncertain travel times and time to deliver an order, gaps appear between route planning and execution of the delivery routes. Additionally, during the planning process additional customers may order and choose new time slots. This fuels the need for data-driven approaches that can account for these uncertainties. In this paper, we present an overview of ORTEC applications that use machine learning combined with route planning, to create accurate, yet efficient and robust route plans. To validate the impact of the proposed approaches in practice, the machine learning models are trained and tested on real-world data. Simulation experiments highlight the impact and value created by such approaches for home delivery companies.