Innovation Platforms (IPs) are seen as a promising vehicle to foster a paradigm shift in agricultural research for development (AR4D). By facilitating interaction, negotiation and collective action between farmers, researchers and other stakeholders, IPs can contribute to more integrated, systemic innovation that is essential for achieving agricultural development impacts. However, successful implementation of IPs requires institutional change within AR4D establishments.

Adapting through innovation is one way for rural communities to sustain and improve their livelihoods and environments. Since the 1980s research and development organizations have developed participatory approaches to foster rural innovation. This paper develops a model, called the Learning-to-Innovate (LTI) model, of four basic processes linked to decision making and learning which regulate rate and quality of innovation. The processes are: creating awareness of new opportunities; deciding to adopt; adapting and changing practice; and learning and selecting.

The three system CGIAR research programs on Integrated Systems for the Humid Tropics, Dryland Systems and Aquatic Agricultural Systems have included “capacity to innovate” as an intermediate development outcome in their respective theories of change. The wording of the intermediate development outcome is “increased systems capacity to innovate and contribute to improved livelihoods of low-income agricultural communities.” This note captures the CGIAR's collective thinking about this intermediate development outcome from a systems perspective to clarify it and inspire other programs.

This paper describes the learning selection approach to enabling innovation that capitalizes on the complexity of social systems at different scales of analysis. The first part of the paper describes the approach and how it can be used to guide the early stages of setting up a “grassroots” innovation process. The second part of the paper looks at how the learn selection model can be used “top-down” to guide research investments to trigger large-scale systemic change.

This paper (Part I) present a case study of work conducted by the International Centre for Tropical Agriculture (CIAT) to adapt network mapping techniques to a rural and developing country context. It reports on work in Colombia to develop a prototype network diagnosis tool for use by service providers who work to strengthen small rural groups. It is complemented by a further paper in this issue by Louise Clark (Part II) which presents work to develop a network diagnosis tool for stakeholders involved in agricultural supply chains in Bolivia.

There have been repeated calls for a ‘new professionalism’ for carrying out agricultural research for development since the 1990s. At the centre of these calls is a recognition that for agricultural research to support the capacities required to face global patterns of change and their implications on rural livelihoods, requires a more systemic, learning focused and reflexive practice that bridges epistemologies and methodologies.

The Great Lakes region of Central Africa is an area abundant in hills, people and conflicts. Its high altitude and cooler climate make it ideal for agriculture. But soils have been exhausted, spare land is no longer available, and farm households in parts of this region rank among the most food insecure and malnourished on earth. Years of civil conflict have moreover paralyzed agricultural advisory and extension services and resulted in poor access to markets.

A platform of farmers, retailers and service providers, civil society organisations, NGOs, government officials, and researchers improves livelihoods in Rwanda. Through interaction and collaboration, these groups experiment with various technological and institutional innovations, thereby tackling local agricultural challenges. This experience illustrates the importance of institutionalising a space where knowledge can be co-created

This case study zooms in on multi-stakeholder processes in the East and Central Africa (ECA) Action Area or Flagship that were launched on 20 May 2013 in Bukavu, DR Congo. The ECA Flagship encom-passes the Rwanda, DR Congo, Burundi, Uganda, Kenya and Ethiopia Action Sites. More specifically, the case study describes and reflects upon the first two years of CGIAR Humidtropics in DR Congo. aiming to outline the multi-stakeholder process as it unfolded and highlight lessons that can be learned from this.

To support the multi-stakeholder process in Burundi, the national research institute ISABU (Institut des Sciences Agronomiques du Bu-rundi) was requested to act as the facilitating organisation. ISABU had previously partnered with the CGIAR centres in Burundi under the CIALCA program that had supported partnerships to coordinate activities and stimulate demand-driven research. With the aim of building on existing collaboration and activities, it was decided to re-engage with former CIALCA partners, including ISABU.