During the past half century, filtration techniques based on the use of activated carbons have been widely used in industrial applications, and considerable knowledge about the mass transfer processes involved has been developed. Surprisingly, the use of activated carbon filters in building applications has remained marginal, and one reason is probably the lack of information on the actual efficiency of such systems in realistic configurations of building operation. The present study addresses this question by investigating the potential of packed-bed activated carbon filters to improve the indoor air quality of residential and office buildings. The method is based on dynamic simulations with a filter model accounting for advective transports, boundary layer diffusion around the activated carbon grains, diffusion and adsorption / desorption within the grains. Particular emphasis has been put on implementing elemental models that suits the specificities of indoor air quality compared to industrial applications. Especially the model accounts for the influence of temperature, humidity and competition for adsorption between all the species contained in the air. The ability of the filter to clean the air is determined by its breakthrough time for various contaminants at typical concentration levels indoors. Considering the number of contaminant found in indoor settings, this shows that the question of the efficiency of the filter cannot be answered globally, and that the question of how often the filtering medium must be changed is not as simple as one can think.