Learning While Implementing: A Reflection from DemocráTICa

By Agustina Ascani

Learning & Knowledge

There is a concept I find particularly useful when reflecting on the Knowledge & Learning process in DemocráTICa: Kaizen, the idea of continuous improvement through iteration.

I do not use it to suggest that DemocráTICa was designed around Kaizen. Rather, it offers a useful lens for describing something that became increasingly evident throughout implementation: the programme was continuously evolving, and our understanding of what it was learning had to evolve with it.

This became particularly clear when we began consolidating the programme’s learning. The initial task seemed simple: identify the main lessons, organize the evidence, and articulate them. In practice, it was much more complex. Learning was not sitting in one place waiting to be collected. It was distributed across mechanisms, teams, actors, conversations, decisions, adaptations, and relationships. Some of it was captured through monitoring data and surveys; some through interviews and qualitative assessments; and some remained embedded in the tacit knowledge of those closest to implementation.

The challenge was therefore not simply to collect learning, but to make it visible — distilling lessons and good practices from a complex initiative into clear, valuable insights that can inform future programmes of similar scale and complexity.

Making learning visible

This meant repeatedly asking: What happened? What does it tell us? And what should we do with that knowledge?

The learning generated throughout DemocráTICa was much broader than any single set of lessons could capture. The ten lessons highlighted here are a selection of insights that help tell a broader story about how the programme was designed, implemented, adapted, and learned from over time.

The ten highlighted lessons can be understood as interconnected rather than isolated:

First, programme architecture and design matter. One of the main learnings was that complementarity works when mechanisms have clear roles. The Fund, Network, and POLIS operated autonomously, each with its own operational design, participation criteria, and tools, while contributing complementarily to the overall programme architecture. Their differentiated functions created value when they were intentionally connected rather than treated as isolated interventions. The experience also showed that time is a critical programme resource: meaningful participation, trust-building, capacity strengthening, and collaboration cannot always be compressed into short implementation cycles. Finally, the co-creation process reinforced the value of diverse voices in programme design and implementation, as different actors brought contextual knowledge and perspectives that helped the programme remain responsive as needs and conditions evolved.

Second, implementation revealed that support needs to go beyond the initial intervention. The Fund demonstrated that financial support can function as an enabler, rather than simply as a transfer of resources. Funding could open pathways to technical, administrative, legal, and programmatic accompaniment, creating conditions for actors to address barriers beyond the grant itself. This was an important distinction from a more transactional model of international cooperation. The experience also showed that support models need to evolve as implementation reveals new needs: Clinics, for example, expanded beyond their initial technology-focused approach toward broader forms of accompaniment. Continuous adaptation is therefore a programme capacity, not simply a response to problems. Collaboration also required more than bringing actors together. The programme surfaced what can be understood as relational latency: actors may already know of one another and even share agendas, but this does not necessarily translate into active collaboration. Relationships need trust, time, intentional and structured spaces, and opportunities for exchange to become meaningful. The learning was therefore not simply that networks matter, but that programmes can help create the conditions for existing relational potential to become collaboration.

Third, learning and evidence require both systems and collective interpretation. The programme had to navigate the tension between reach and depth: engaging a broad ecosystem while also providing more sustained support to a smaller number of actors. This required deliberate choices about where depth could generate the most value and how insights from that work could inform the wider programme. It also reinforced the importance of designing Monitoring and Evaluation (M&E) and Knowledge & Learning (K&L) systems from the outset, creating a foundation for evidence to accumulate throughout implementation. The programme operated within a model of distributed governance, where leadership and decision-making were shared across partners, mechanisms, and teams rather than concentrated in a single structure. In this model, learning also had to move across levels, informing decisions beyond the point where it was generated. Such a model required active orchestration to connect perspectives and maintain coherence while preserving the autonomy and differentiated roles of each part of the programme.

Finally, how value and scale can be understood. Value is not limited to immediate outputs or to what happens within an individual intervention. The programme generated forms of capacity that remained, knowledge that circulated, methodologies that were adapted, and relationships that created new opportunities. Looking at these dimensions makes it possible to ask not only what an intervention delivered, but what it enabled beyond the initial point of support. Scale is not only about reaching more actors; it can also mean creating conditions to persist and travel across the ecosystem.

Looking back at the process of making these lessons visible, what stands out is that learning did not happen after implementation — it happened through implementation.

The process therefore involved more than documenting what happened. It meant distinguishing individual experiences from broader programme patterns, and outputs from insights that could be meaningful beyond a specific context. Not every learning needed to become a recommendation; some were valuable because they challenged assumptions, surfaced questions, or showed what future programmes may need to pay attention to earlier. For K&L, the work was ultimately about moving from experience to patterns, and from patterns to insights that could be used beyond the programme itself.

Building for learning

This process also highlighted that learning needs to remain connected to implementation rather than becoming a separate exercise at the end of a programme. As evidence accumulated, K&L had to help connect emerging insights with the questions, decisions, and adaptations already taking place. This required creating space not only to document what had happened, but to revisit assumptions, test emerging interpretations, and identify which insights could be relevant beyond a specific experience.

This is perhaps where the idea of Kaizen becomes most useful. Continuous improvement depends not only on making the next iteration better, but on retaining and making use of what was learned in the previous one. For programmes of this scale and complexity, learning becomes part of what the programme leaves behind — not as a fixed blueprint, but as a set of tested insights, practices, and questions that can inform future action.

Thanks for reading.