Our aim was to demonstrate that an understanding of the process of how research may (or may not) influence policy and practice is crucial to leverage research findings and bring about evidence-informed policy and its implementation. We describe a process of research design and execution, based on theories of the relationship between evidence and public policy-making, which sought to improve the uptake of evidence into the HIV policy-making process in Pakistan. We designed and implemented specific strategies in research methods, management and dissemination to increase the policy influence by recommendations from a multi-disciplinary research project. Research to policy is complex, rarely linear and causal attribution is problematic. Nonetheless, we believe that, in part, some of the current changes in HIV policy and practice in Pakistan may be due to the managed process of research influence. We offer four key recommendations for those concerned with improving the chances of seeing their research incorporated into policy and practice – these are (1) involve stakeholders in research management; (2) set realistic expectations of research impact; (3) invest in long-term research–policy-maker relationships; and (4) build capacity of end users to use research to demand policy change. [ABSTRACT FROM AUTHOR]