Transdisciplinarity connotes a strategy that crosses many disciplinary boundaries to create a holistic approach. Due to this insistence, it has gained widespread popularity in recent years. However, in transdisciplinary collaborations based on academic–practitioner interactions, this is not always as straightforward. In this text, we share some insights from our past and ongoing work with the project ‘Urban Rural Gothenburg’, within which we have launched the Research Forum (RF) model as a means of co-producing new transdisciplinary knowledge. RF ‘Urban Rural Gothenburg’ constitutes Mistra Urban Futures' contribution to the project ‘Urban Rural Gothenburg’, a three-year (2017–19) EU-sponsored project for sustainable development with the overarching aim to create improved conditions for green innovation and green business development between the city and the countryside. The RF constitute the project’s academic component within a transdisciplinary (penta-helix) model. The RF is meant to serve as an incubator and accelerator of various initiatives concerned with understanding, testing and implementing ecologically oriented solutions that may arise through academic–practitioner interactions. The RF is thus not a ‘place’ (in the concrete sense) but a collaborative effort of two coordinators – one practitioner and one academic, aided by an assistant, who actively pursue and facilitate new ways of extracting knowledge within a large and heterogenous project structure. Identifying and successfully matching different perspectives, points of view and pools of knowledge is a difficult challenge. This is mainly because interactions are seldom based on the same principles; different people have different foci, incentives, and agendas, while understanding how they work out in practice is key to successful implementation of the RF model. In this presentation, we focus on the description, analysis and evaluation of the RF as a methodological endeavor. The findings center on four of the most common modes of interaction encountered during our work with the RF: academics to practitioners (A > P); practitioners to academics (A < P); academics with practitioners (A >< P); and academics without practitioners (A | P). We conclude that if we truly want to embrace co-production as way to obtain new knowledge we inherently must concede part of our individuality towards a homogenous goal. At the same time, the specificity of different forms of knowledge cannot be melted into an amorphous mass, elsewise co-production is likely to become a tokenistic effort of little applicatory utility. Put simply, we must constantly remain open to change but also stay protective of knowledge that works without reinvigoration.