In late 2025, the Dgroups Foundation hosted a virtual Knowledge and Learning Exchange with practitioners to reflect on online collaboration, knowledge sharing, and community practice.
The session was co-hosted by Dgroups Foundation Associates Jessica Ball and Pier Andrea Pirani, with contributions from Rocío Sanz, Davide Piga, Simon Hearn, and Ivan Piseta, alongside participants joining from a wide range of organisations and regions.
The exchange was structured around four themes:
- Knowledge retention;
- The role of AI in knowledge management and collaboration;
- How to understand and assess success in online collaboration and knowledge exchange; and
- What it takes to sustain and grow engagement over time.
These themes provided a shared frame for the conversation.
Why knowledge retention still matters
Rocío opened by emphasising that knowledge retention goes beyond documents and databases. While reports and systems capture part of what organisations know, much of the most valuable knowledge remains tacit, shaped by experience, context, and relationships.
“You may leave a report behind, but you take with you the ‘why’, the ‘how’, and the ‘who’.”
Participants reflected on how short contracts, staff turnover, and fragmented ways of working often lead to the quiet loss of institutional memory. Knowing where organisations and communities come from is not about repeating the past, but being better equipped to respond to the future, informed by what has already been tried, learned, and adapted.
Simon added that communities of practice offer an important space for this kind of learning, bringing together participation (conversation and interaction) and reification (documents, tools, and artefacts) in ways that allow knowledge to be interpreted and reused, not just stored.
AI as a support, not a substitute
A second focus was on the roles of artificial intelligence in knowledge management and collaboration.
Davide shared practical examples from the recent relaunch of the KM4Dev Knowledge Sharing Toolkit, where AI is used to reduce barriers to contribution. Instead of asking contributors to produce polished written content, the toolkit allows people to talk through a method or experience, with AI helping to structure this into a first draft that follows agreed templates and style.
“AI helps people get started. Humans need to keep the last mile: judgment, meaning, and quality.”
AI is also being used to support reflection, helping practitioners surface lessons learned from experience and link those insights to existing methods. This approach aimed to keep people at the centre while making contributions easier.
At the same time, speakers and participants were clear about limitations. AI produces fluent language without understanding context or truth, and can introduce unnecessary complexity if not used carefully. Several contributors stressed that AI should support thinking and reflection – not replace them – and that tacit knowledge still depends on human interaction.
Ivan added that, from a platform perspective, one promising direction is using AI to improve discoverability, helping users find relevant content more easily, rather than asking AI to generate new conclusions or advice.
Collaboration and knowledge exchange are not projects
The third theme was how to understand and assess “success” in online collaboration and knowledge exchange, particularly in settings where participation is voluntary, and outcomes are not easily reduced to metrics.
Simon offered a framing that resonated strongly: online collaboration spaces and communities cannot be treated in the same way as projects or programmes. Participants join for different reasons, and conveners, host organisations, and funders often bring their own expectations.
“You can‘t impose a vision of success on a community. It has to be negotiated and shared.“
Simon distinguished between internal outcomes, such as changes in participation and collaboration practices, and external outcomes, where learning influences work beyond the collaboration space itself. He also noted that conventional evaluation tools, such as surveys, often struggle to gain traction in voluntary contexts.
Reflection and learning, he suggested, are more effective when embedded into activities that already matter to participants, supporting learning rather than accountability alone.
Onboarding, trust, and sustaining engagement
The final theme focused on sustaining and growing engagement over time, bringing the discussion back to fundamentals that cut across all four areas.
Rocío highlighted how the shift to online work has reduced informal spaces for connection and sense-making. Well-designed collaboration spaces can help recreate some of this, but only if people feel welcomed and safe to contribute.
Davide emphasised that onboarding is not a technical step but a relational one.
“Onboarding is how people learn what kind of room they’ve entered and whether it’s safe to speak.”
Clear expectations, shared values, and explicit norms help participants engage without fear of judgment or reputational risk. Several speakers noted that small, intentional practices (e.g., welcome messages, community charters, etc.) often have a decisive impact on engagement.
Simon offered a reminder that most participants will never post actively, and that this is not necessarily a problem.
“The role of conveners is often to keep the campfire going at the core.”
By supporting a small group of active contributors, collaboration spaces can sustain energy while still creating value for quieter participants.
Shared value and looking ahead
Looking across the discussion threads, people highlighted the importance of shared value, not only among participants, but also between collaboration spaces and the organisations that support them. For engagement to be sustainable, participation needs to be meaningful for individuals and aligned with organisational priorities.
Despite different roles and contexts, participants described similar challenges and lessons. Technologies will continue to evolve, and funding environments will remain uncertain. Yet many fundamentals remain constant:
Trust takes intention. Onboarding matters. Stories carry knowledge. AI can support, but people make meaning.