Poor farmers are particularly vulnerable to environmental stressors and often rely on traditional knowledge and grassroots creativity to help them to adapt. Such adaptation can be enhanced using other knowledge sources, but this requires greater understanding of the processes of knowledge co-production among scientists, local communities and state actors. In this study we undertook knowledge co-production on an experimental basis with two contrasting communities: smallholder farmers in Jind (Haryana) and Adi women in East Siang district (Arunachal Pradesh). We found that the Jind farmers displayed grassroots creativity in coping with salinity induced stresses to rice-wheat cropping systems, while Adi women applied their traditional knowledge of food based on namdung (Perilla ocymoides, a local plant species) to cope with climate variability that affected fermentation. Jind farmers perceived the process of knowledge co-production as moderately credible and salient, but the legitimacy of the exercise was compromised by the relatively low level of participation by state actors. The farmers rated the practical outcomes of the co-produced zero-till wheat as low to moderate for combating salinity-induce risks but as high for a community rice nursery. The knowledge co-production process was considered more credible and salient among younger Adi women than older women, and the utility of the Adi women's co-produced adaptive practices were rated as moderate to high for reducing the impacts of climate variability on namdung based foods. In both cases, an emergent property of the knowledge co-production exercise was creation of a knowledge network that has the potential to lead to ongoing enhanced adaptation to environmental change. Insights from the study could help improve knowledge co-production in similar social-ecological systems, and can be integrated with environmental change policies.