Knowledge is a critical enabling factor for healthy agri-food innovation systems (AIS). AIS and related knowledge management (KM) frameworks face significant implementation challenges. We review applications of KM to AIS, the current state of the art and shortcomings and present a new KM framework, Agricultural Knowledge Management for Innovation (AKM4I). Previous agricultural KM frameworks do not integrate innovation pragmatically, use linear, reductionist, top-down pathways to innovation, and do not explicitly incorporate issues of power, politics, ownership, and trust when combining scientific and local knowledge across multiple stakeholders. The AKM4I framework addresses systemic interactions favouring innovation outcomes by formalising flows and management of information and knowledge between diverse sets of stakeholders; and explicitly considering previously unresolved practical and relational barriers aiming to facilitate more equitable, rapidly evolving, and actionable knowledge generation and management for innovation and transformational change. An agricultural case study serves as an example of the implementation of AKM4I.

Agri-food systems are besieged by malnutrition, yield gaps, and climate vulnerability, but integrated, research-based responses in public policy, agricultural, value chains, and finance are constrained by short-termism and zero sum thinking. As they respond to current and emerging agri-food system challenges, decision makers need new tools that steer toward multi-sector, evidence-based collaboration. To support national agri-food system policy processes, the Integrated Agri-food System Initiative (IASI) methodology was developed and validated through case studies in Mexico and Colombia. This holistic, multi-sector methodology builds on diverse existing data resources and leverages situation analysis, modeled predictions, and scenarios to synchronize public and private action at the national level toward sustainable, equitable, and inclusive agri-food systems. Culminating in collectively agreed strategies and multi-partner tactical plans, the IASI methodology enabled a multi-level systems approach by mobilizing design thinking to foster mindset shifts and stakeholder consensus on sustainable and scalable innovations that respond to real-time dynamics in complex agri-food systems. To build capacity for these types of integrated, context-specific approaches, greater investment is needed in supportive international institutions that function as trusted in-region ‘innovation brokers.’ This paper calls for a structured global network to advance adaptation and evolution of essential tools like the IASI methodology in support of the One CGIAR mandate and in service of positive agri-food systems transformation.

Traditional agricultural extension services rely on extension workers, especially in countries with large agricultural areas. In order to increase adoption of sustainable agriculture, the recommendations given by such services must be adapted to local conditions and be provided in a timely manner. The AgroTutor mobile application was built to provide highly specific and timely agricultural recommendations to farmers across Mexico and complement the work of extension agents. At the same time, AgroTutor provides direct contributions to the United Nations Sustainable Development Goals, either by advancing their implementation or providing local data systems to measure and monitor specific indicators such as the proportion of agricultural area under productive and sustainable agriculture. The application is freely available and allows farmers to geo-locate and register plots and the crops grown there, using the phone’s built-in GPS, or alternatively, on top of very high-resolution imagery. Once a crop and some basic data such as planting date and cultivar type have been registered, the application provides targeted information such as weather, potential and historical yield, financial benchmarking information, data-driven recommendations, and commodity price forecasts. Farmers are also encouraged to contribute in-situ information, e.g., soils, management, and yield data. The information can then be used by crop models, which, in turn, send tailored results back to the farmers. Initial feedback from farmers and extension agents has already improved some of the application’s characteristics. More enhancements are planned for inclusion in the future to increase the application’s function as a decision support tool.

Future societal systems will be characterized by heterogeneous human behaviors and data-driven collective action. Complexity will arise as a consequence of the 5th Industrial Revolution and 2nd Data Revolution possible, thanks to a new generation of digital systems and the Metaverse. These technologies will enable new computational methods to tackle inequality while preserving individual rights and self-development. In this context, we do not only need data innovation and computational science, but also new forms of digital policy and governance. The emerging fragility or robustness of the system will depend on how complexity and governance are developed. Through data, humanity has been able to study a number of multi-scale systems from biological to migratory. Multi-scale governance is the new paradigm that feeds the Data Revolution in a world that would be highly digitalized. In the social dimension, we will encounter meta-populations sharing economy and human values. In the temporal dimension, we still need to make all real-time response, evaluation, and mitigation systems a standard integrated system into policy and governance to build up a resilient digital society. Top-down governance is not sufficient to manage all the complexities and exploit all the data available. Coordinating top-down agencies with bottom-up digital platforms will be the design principle. Digital platforms have to be built on top of data innovation and implement Artificial Intelligence (AI)-driven systems to connect, compute, collaborate, and curate data to implement data-driven policy for sustainable development based on Collective Intelligence.

No abstract