Innovation policy and business strategy often expect that investing in private and public research and development will immediately produce a flow of products and processes with high commercial and social returns. Policymakers and managers implicitly follow the logic underlying most linear innovation models assuming a well-defined and uni-directional relationship between R&D spending as input and innovation rents as output of the innovation process. Modern innovation economics dismisses the simplified approximation of knowledge by R&D investment and, instead, considers complex knowledge generation and diffusion processes in innovation networks. From this angle, the disappointing performance of traditional approaches is traced back to strong limits of conventional steering, control, and policy instruments. In this paper, we show that the new view of knowledge generation and diffusion in innovation networks allows for an alternative and has led to systemic approaches in innovation analyses. Combined with computational approaches like agent-based modeling, this new view enables today innovative tools in policy consulting. Using the example of regional innovation policy, we introduce a policy laboratory in which innovation processes can be analyzed in depth to see the impact of different innovation policy instruments in-silico. This ex-ante evaluation helps considerably to improve the understanding of innovation processes and with it the performance of innovation policy.