In one embodiment, example systems and methods relate to ordering food from a vehicle. A driver or a passenger of a vehicle indicates that they would like to order food from the vehicle by speaking a command, or by making a selection on a display associated with the vehicle. In response to the indication, the vehicle determines one or more restaurants that are near the vehicle, or that are near the route being traveled by the vehicle. After the driver or passenger selects a restaurant, the vehicle loads a restaurant specific natural language understanding model that is associated with the restaurant. As the driver or passenger speaks their order, the restaurant specific natural language understanding model is used to process the audio and generate an order. The order and an estimated time of arrival of the vehicle is provided electronically to the restaurant.