There is keen interest in many jurisdictions in finding ways to improve the way that research evidence informs policy. One possible mechanism for this is to embed academics within government agencies either as advisers or full staff members. Our commentary argues that, in addition to considering the role of academics in government as proposed by Glied and colleagues, we need to understand better how research and policy interactions function across policy sectors. We believe more comparative research is needed to understand if and why academics from certain disciplines are more likely to be recruited to work in some policy sectors rather than others. We caution against treating government as monolithic by advocating the same model for collaborative interaction between academics and government. Lastly, we contend that contextualized research is needed to illuminate important drivers of research and policy interactions before we can recommend what is likely to be more and less effective in different policy sectors.