This paper studies innovation in the creative industry of art restoration, which is characterised by an intensive use of symbolic knowledge. Using the resource-based view of the firm as a theoretical framework, this study adapts Community Innovation Survey (CIS) methodology to this industry, creating and exploiting a unique dataset from the restoration departments of museums in 43 countries on 5 continents. The results suggest that the type and composition of the knowledge bases in play influence a department's absorptive capacity to access external information sources and thereby impact innovative outcomes. The article contributes to innovation literature by capturing innovation processes in a symbolic-based industry.