Tuberculosis and HIV/AIDS have synergistic health impacts in terms of disease development and progression. Therefore, collaborative TB and HIV/AIDS activities are a logical health systems response. However, the establishment of these activities presents a challenge for countries that have strong vertical disease programs that differ in their implementation philosophies. Here, we review the process by which TB/HIV collaboration was established in Cambodia. A cycle of overlapping and mutually reinforcing initiatives - local research; piloted implementation with multiple options; and several rounds of policy formulation guided by a cross-functional Technical Working Group - was used to drive nationwide introduction of a full set of TB/HIV collaborative activities. Senior Ministry of Health officials and partner organizations brought early attention to TB/HIV. Both national programs implemented initial screening and testing interventions, even in the absence of a detailed, overarching framework. The use of multiple options for HIV testing identified which programmatic options worked best, and early implementation and pilots determined what unanswered questions required further research. Local conduct of this research - on co-treatment timing and TB symptom screening - speeded adoption of the results into policy guidance, and clarified the relative roles of the two programs. Roll-out is continuing, and results for a variety of key indicators, including screening PLHIV for TB, and testing TB patients for HIV, are at 70-80% and climbing. This experience in Cambodia illustrates the influence of health research on policy, and demonstrates that clear policy guidance, the pursuit of incremental advances, and the use of different approaches to generate evidence can overcome structural barriers to change and bring direct benefits to patients.