The gardening metaphor is precise. A gardener does not design a plant; she creates conditions under which the plant's own growth processes can unfold — the right soil, light, water, protection from elements that would destroy the seedling before roots establish. Communities of practice cannot be designed top-down; their character, repertoire, and identity are emergent properties that resist specification. But they can be cultivated — provided with conditions favorable to emergence, supported with resources that sustain development, protected from organizational pressures that would crush them before they mature.
Five conditions must be cultivated for communities to flourish in the AI age. Shared problems requiring collective engagement: reserve certain decisions for participatory deliberation even when AI could resolve them individually. Maintained legitimate peripheral participation: design alternative peripheries that provide formative experiences for newcomers when traditional peripheral tasks have been automated. Protected spaces for participatory meaning-making: defend time and structure for conversations that don't produce optimizable outputs. Cultivated brokers: develop integrators whose multimembership produces constellation-level perspective. Dual nature of practice: ensure reifications are continually re-engaged through participatory processes.
The institutional investments emerging in 2024-2025 — the federal AI Community of Practice, Columbia's AI community, Harvard's Digital Data Design Institute — represent early cultivation attempts. Their success depends less on the technology they address than on the social structures they build around it. The same patterns that distinguish deep communities of practice from shallow networks — sustained mutual engagement, identity formation, developed shared repertoire — determine whether the cultivated communities will generate genuine social learning.
The challenge the framework names is that cultivation must be deliberate in environments where it was once automatic. The stand-up meeting Segal almost canceled, the breakfast conversations the Xerox technicians had, the code review that caught not just bugs but patterns of thinking — these used to happen because the work happened. In AI-augmented work, they must be designed for, protected, invested in against the constant pressure to optimize them away.
Wenger, Richard McDermott, and William Snyder published Cultivating Communities of Practice in 2002, developing the prescriptive extension of the 1998 theoretical framework for organizational practitioners. The book drew on cases from companies including Shell, Ford, IBM, and the World Bank, documenting how communities of practice had been deliberately supported in settings where they might not have emerged spontaneously.
The framework was extended further in Digital Habitats (2009, with Nancy White and John D. Smith), which addressed cultivation in digital and distributed environments. The AI age represents the most consequential test yet of whether deliberate cultivation can substitute for the organic conditions that Wenger's original theory assumed.
Gardening metaphor. Create conditions for emergence, not blueprints for construction.
Five conditions to cultivate. Shared problems, maintained periphery, protected spaces, cultivated brokers, dual nature of practice.
Runs against efficiency pressure. The spaces required for community formation look wasteful by productivity metrics.
Requires institutional commitment. Cultivation fails when organizations optimize away the conditions it depends on.
Substitutes for organic emergence. Essential in environments where the conditions for spontaneous community formation have been eliminated.