Digital transformation is primarily an institutional and cultural challenge, not a technological one. Across national and local governments, fragmentation, weak coordination, and limited digital literacy are bigger barriers than lack of tools. Combining top-down alignment with bottom-up experimentation is essential for sustainable digital transformation. Institutional strategies gain traction only when they are complemented by locally co-created solutions grounded in real user needs. Capacity-building must go beyond skills transfer and address trust, incentives, and mindsets. Training is most effective when it is embedded in real projects and linked to decision-making processes. People-centered and participatory methods reveal blind spots invisible to traditional diagnostics. Interviews, workshops, and system mapping uncovered informal practices, workarounds, and risks (e.g., data misuse, exclusion) that are not captured in formal assessments. Responsible use of emerging technologies requires governance frameworks and ethical foresight from the outset. Without this, tools such as AI or data platforms risk reinforcing inequalities or creating new forms of exclusion.

-

These data sources addressed gaps in traditional diagnostics by capturing lived experiences, informal coordination practices, levels of digital literacy, trust dynamics, and institutional incentives that are often invisible in quantitative or policy-level data. Digitalization efforts reveal significant data gaps within public administrations, including limited visibility over internal processes, overlapping functions, and the lack of systematized data. Fragmented communication and grievance redress channels—often supported by outdated technologies—further constrain the effective use of information. These gaps limit both efficient service delivery and the ability to understand population needs and emerging demands. A new line of work different from the Buenos Aires province and Comodoro Rivadavia City was related to IA and digital capacity building at the parliamentary level. The Argentine Senate requested UNDP’s support to inform its AI strategy. In response, we conducted a participatory workshop to assess the current state of knowledge and use of AI among Senate staff. The process revealed significant information gaps, including uneven understanding of AI capabilities and limitations, lack of shared criteria for responsible use, limited access to practical guidance tailored to parliamentary work, and insufficient training opportunities. These gaps constrained informed decision-making and the consistent adoption of AI tools across teams.

These methods allowed us to unpack the system beyond formal structures, identify root causes of fragmentation, test assumptions early, and explore long-term implications of digital transformation. The interplay of innovation methods, new forms of data, and unusual partners enabled learning that would not have been possible through administrative data alone. In the case of the Ministry of Community Development of Buenos Aires Province, fragmented records and non-interoperable systems meant that core processes—such as how citizen requests were received, routed, and resolved—were largely invisible at an aggregate level. Through service design and co-creation, frontline staff were engaged as key knowledge holders, revealing overlaps between areas, informal coordination practices, and recurring demand patterns that were not captured in any database but were central to how the system actually worked. As in the collaboration with the Senate, Vvalidation workshops and process mapping transformed this tacit, experiential knowledge into structured insights, effectively generating new data where none previously existed. In the case of Comodoro Rivadavia, innovation emerged from the collective interpretation of partial, contextual, and lived information rather than from a single authoritative source. Behavioural science helped explain why staff routinely went beyond their formal mandates to respond to citizen demands, while collective intelligence methods aggregated individual observations into shared diagnoses of institutional strain. Prototyping and iterative feedback loops enabled rapid learning by testing assumptions in practice, and strategic foresight situated current bottlenecks within longer-term trends—such as rising demand and evolving digital expectations—highlighting future data needs. Together, these approaches expanded the evidentiary base and enabled the generation of actionable insights precisely in areas where data gaps were most acute.