As Roger Martin has noted, organizations often confuse strategy with planning. Workplans with timelines and budgets are artifacts of certainty. They give leaders the comfort of feeling in control while avoiding difficult choices and tradeoffs.
Those trade-offs are the domain of strategy, and the public sector is no exception.
So when RIL – Red de Innovación Local, a South American nonprofit that connects more than 10,000 public servants across 800 cities in 30 countries, set out to improve how municipalities build strategy, its founder, Delfina Irazusta, knew the challenge would not be technical; it would be cultural.
To tackle it, she took the old idiom “eat your own dog food” to heart. She decided that her team would first try out a new strategy process for itself before bringing it to municipalities. This way, they would have the street credentials from having done the work, and they would be in a better position to provide advice on implementation.
The experiment centered on PortalRIL, RIL’s AI-powered platform trained on over a decade of local government knowledge. But as Stefan Verhulst has observed, “The most consequential failures in data-driven policymaking and AI deployment often stem not from poor models or inadequate datasets but from poorly framed questions.”
Starting With Inquiry
It is for this reason that RIL team members were asked to start their strategy process in a rather unconventional manner, through a “Questions Tree,” or a structured reflection process guided by questions focusing on 3 different levels:
- Organisation-wide issues
- Individual questions on each member’s role
- Team level questions.
This team exercise produced a collaborative document summarizing the key questions informing RIL’s strategy going forward.
In the second phase of the process, "Diagnostics," team members were asked to rely on PortalRIL for a retrospective exercise.
Individually, they reflected on questions such as “what worked well in the last year in my line of business, what didn’t? And what should I do/more or less of?” Participants then prepared a short document, which they would refine and summarise using AI.
Initial skepticism quickly gave way to surprise:
“At first, I was unsure about how AI could effectively fit into the daily realities of our administration. However, I was surprised by its capacity to understand the municipal context and provide precise solutions. It proved that AI isn’t just a tech trend; it’s a practical, powerful ally for improving public management.”
From Silos to Systemic Thinking
In the third phase of “Connections,” PortalRIL analyzed the collective reflections to identify patterns, such as what challenges were shared? And where are potential synergies to explore?
One participant captured the shift:
“The suggestions were spot on and offered ideas I hadn’t considered favorable for administration tasks before. It helped me systematize my thoughts incredibly fast, offering a variety of justified options I wouldn’t have reached on my own.”
What might once have required months of cross-team coordination emerged quickly as a shared mental model. AI functioned as a radar, surfacing patterns and connections, enabling team members to slowly form a common mental model of areas worth focusing on.
Using this framework, the team entered "Ideation" mode with the help of AI. Here, the Questions Tree again served as a device to anchor suggestions, distinguishing among the organisation, team, and individual levels. This phase ended with a long list of divergent ideas generated using PortalRIL.
The last phase is where AI-generated inputs meet the collective intelligence of the group: team members were asked to confront the pros and cons of different ideas and evaluate trade-offs. When they got stuck, they were encouraged to use PortalRIL to generate more ideas.
Another insight stood out:
“What I value most is how it connects information from other governments. It serves as a bridge to investigate and relate successful experiences from other places to our own work. It allows us to compare management styles and learn from peers in a way we couldn’t before.”
Broad strategic directions emerged from the ideation process and were translated into actionable insights during the design phase. Here again, AI was used to support the transition from strategic intent to a portfolio of actionable options, which were then consolidated in a final phase into a strategic document.
What This Experiment Reveals
Government capacity is built through doing, and through its learning journey, the RIL team is now in a stronger position to advise cities on how to reimagine their strategy process. Three key insights emerged from their experiment:
- The Transformative Power of Inquiry: By prioritizing the "Questions Tree" as the starting point, the team moved beyond conventional problem-solving to a profound understanding of underlying challenges. This phase was critical not just for strategy, but for defining the organization's identity in the AI age.
It triggered existential reflections—such as questioning what unique role facilitators should play to augment AI-surfaced patterns—which shifted the focus from operational efficiency to human value. This ensured that the strategy was built on robust foundations, prioritizing empathy and leadership alongside technical solutions.
2. Expanding Strategic Horizons: The integration of PortalRIL significantly broadened the spectrum of options, revealing opportunities to add value that traditional methods might have missed. A clear example was the exploration of using AI for massive personalization, such as automatically adapting project modules to the specific language and format of different donors.
This opened new pathways for scalability and prompted the team to delegate operational tasks to AI agents, freeing up capacity to concentrate on creativity and strategic reasoning.
3. Accelerating Synergies and Systemic Speed: Perhaps the most tangible breakthrough was the speed at which AI bridged organizational silos. Acting as a systemic radar, it processed collective inputs to instantly identify intersections that typically require significant coordination.
The AI surfaced natural synergies, and in the case of RIL, this looked like connecting entrepreneurial cities with fleet management or integrating environmental goals with mobility planning. Allowing the team to visualize a synchronized, multidisciplinary roadmap.
This capability transformed what would have been months of isolated planning into an immediate recognition of shared challenges and cross-pollination opportunities.
Next Steps: Getting Cities in Argentina (and Beyond) to Embrace AI-Enhanced Strategy
Building on the successful internal pilot, early indicators suggest a growing embrace of AI-enhanced strategy within municipalities. Cities are increasingly recognizing the potential of tools like PortalRIL to navigate complex challenges, move beyond reactive problem-solving, and foster more data-informed strategic planning.
While the full scope of adoption is still emerging, AI's ability to surface patterns, identify efficiencies, and broaden strategic horizons is proving compelling.
However, RIL’s pilot highlighted that to truly unlock this value, the team must work with municipalities to address a dual challenge found in the territory of bridging the digital divide for cities with lower technological maturity, and ensuring that AI standardization respects and preserves the unique "singularity" and identity of each local government.
RIL is now actively working on diffusing this AI-enhanced strategic planning methodology to cities across the region.
To accelerate this, RIL is launching specific innovation challenges to co-design, alongside cities, a gradual adoption path through small, high-impact actions. The strategy involves comprehensive training programs and direct technical assistance, leveraging PortalRIL as a central platform.
The ultimate goal is to prove that AI is not just about saving operational time, but about reclaiming that time to focus on what truly matters: a deeper connection with citizens, strategic depth, and the actual implementation of policies.