Knowledge dissemination interventions (KDIs) are integral to knowledge brokerage activities in research as part of the ethics of practice, but are seldom evaluated. In this case study, we critically reflect on an annual KDI as part of knowledge brokerage activities in the MRC/Wits-Agincourt Unit health and demographic surveillance system (HDSS) in rural South Africa from 2001 to 2015. The HDSS findings on births, deaths and migrations, as well as nested research project results, were shared with villagers, village leaders and service providers. The data used for this case study comprised secondary analysis of 13 reports and 762 evaluation forms of annual village-based meetings; records of requests for data from stakeholders; and qualitative analysis of 15 individual and five focus group interviews with local leaders and service providers involving 60 people. Over time, the KDI evolved from taking place over one week a year to being extended over six months, and to include briefings with service providers and local leaders. Attendance at village-level meetings remained low at an average of 3 per cent of the total adult population. Since 2011, the KDI village-based meetings have developed into an embedded community forum for discussion of topical village issues. There has been a decrease in requests for health-care and other services from the research unit, with a concurrent increase in research-related questions and requests for data from service providers, village leaders and political representatives. We conclude that, in this setting, the dissemination of research findings is not a linear exchange of information from the researchers to village residents and their leadership, but is increasingly multi-directional. KDIs are a key component of knowledge brokerage activities and involve, influence and are influenced by other aspects of knowledge brokerage, such as identifying, engaging and connecting with stakeholders and supporting sustainability.