In policy consultation, communicating model results to administration and policy makers has always proven to be a challenge for scientists. Many of the relevant preconditions for effective and successful policy advice are aggravated when results are based on the simultaneous use of a multitude of different models. This paper identifies key issues - e.g., relations to administration; correct identification of prevailing objectives of all agents involved; ability to run scenarios ‘in time’ - and discusses strategies for successful communication based on the experiences of the vTI model network. Specific attention is paid to the issue of communicating ‘conflicting’ results of different models: while often seen as a source for scientific insight, such ‘inconsistencies’ have proven to be a major obstacle for acceptance in a non-academic institutional setting. The experiences, as well as the literature, point specifically to the importance of tight linkages between modellers and policy makers, and the need to abandon decisionist or technocratic approaches of policy advice in favour of pragmatic approaches stressing the bilateral nature of communication.