Recently, ridesharing services have grown rapidly. In future, fleets of shared and pooled autonomous vehicles may transform urban mobility. In this paper, we introduce an approach to dynamically simulate these services within a fullstack transport simulation using an insertion-based algorithm. In a first test case, using a taxi data set from Berlin, the potential for shared rides is evaluated using a fleet of vehicles with a capacity between two and four ride requests. The simulation suggests that the overall vehicle kilometers traveled may be reduced by 15-20 % , while travel time increases can be kept at a relatively low level of less than three minutes per person. Additionally, the simulation results suggest in which areas of the city it may be the most rewarding to offer shared services.