Product bundling is a marketing strategy that concerns offering several products for sale as one combined product. Current technology mainly focuses on the creation of static bundles, which involves pre-computing product bundles and associated discounts. However, due to the inherent dynamism and constant change of current and potential customer information, as is particularly the case in enterprise networks, static product bundles prove to be inefficient. In this paper an approach for dynamic generation of personalized product bundles using agents is proposed. Our approach involves creating bundles based on substitution and complementarity associations between product items, and subsequently ranking the produced bundles according to individual preferences and history of each customer. The proposed approach has been implemented in e-Furniture, an agent-based system supporting networking of furniture and wood product manufacturing enterprises.