This study identifies and empirically validates a microfoundational mechanism based on the Intelligent Knowledge Management Model (MIGC) to explain how knowledge systems mediated by artificial intelligence (AI) enable the twin transition in artisanal production contexts. Although the literature has extensively analyzed the relationship between digitalization and sustainability, a critical gap persists in the understanding of the microfoundational processes through which these transformations emerge in environments of low technological intensity and strong cultural roots.
Through a combined approach of design science and participatory action research, an intervention in a rural ceramics system in Guanajuato, Mexico, is implemented and evaluated. The model integrates AI, energy monitoring and artisanal knowledge within an architecture aimed at the recombination of knowledge and the development of organizational capacities. The data were collected over 12 months through energy audits, production records and qualitative fieldwork.
The results show a 19.6% reduction in energy intensity and a 14.2% reduction in production defects, along with improvements in process stability, traceability and decision-making. Beyond these results, the study identifies a recursive microfoundational mechanism in which AI does not act as an automation tool, but as a cognitive infrastructure that facilitates the recombination of tacit and explicit knowledge, activating the reconfiguration of organizational capacities and generating sustainable results.
This finding allows us to reconceptualize the twin transition as a socio-cognitive process based on the articulation of knowledge, rather than as a direct consequence of technological adoption. Likewise, the study demonstrates that dynamics of innovation ecosystems can emerge at microterritorial scales and that their effectiveness depends on the interaction between technology, agency and knowledge structures.
Overall, the results establish that sustainability in contexts of low technological intensity does not depend on intensive automation, but on the design of knowledge architectures that integrate artificial intelligence, local knowledge and organizational agency, opening new lines of research on the scalability of these processes.