CIIAR – Coalition of Cities for Artificial Intelligence

The Problem

Local governments in Latin America faced three critical, interconnected barriers to adopting AI in public administration:

1. Diagnostic gap: Most municipalities had no clear picture of their own digital and data maturity, and no structured starting point for AI adoption.
2. Access to solutions: Despite a growing GovTech ecosystem developing AI-powered tools, there was no efficient mechanism for municipalities to identify, evaluate, and procure solutions proven to work at their scale and context.
3. Peer learning deficit: Every city was tackling the same challenges in isolation — staff turnover, limited budgets, political resistance, and emerging regulatory frameworks — with no structured network for public officials to share real implementation lessons.

Without an accessible entry point and a trusted peer network, AI adoption remained concentrated in large capital cities, deepening the digital divide between cities of different sizes and resources.

The Solution

CIIAR, coordinated by the Red de Innovación Local (RIL) with support from BIDLab, built a three-component AI-enabling ecosystem for local governments across Latin America:

Component 1 – AI Maturity Diagnostic (Autodiagnóstico IA): A free self-assessment tool that allows any municipality to evaluate its readiness across four dimensions: data governance, technology infrastructure, human capacity, and ethical/regulatory frameworks. The tool produces a concrete, personalised roadmap for AI adoption that reflects the city's actual starting point rather than an idealised model.

Component 2 – GovTech Connect: A structured programme that identifies real public management challenges within member cities and runs an open call for startups to develop or adapt AI-powered solutions. Selected solutions are piloted in member cities, generating real-world evidence. "Demo Days" bring municipal officials and govtechs together for hands-on validation and cross-city knowledge transfer.

Component 3 – Regional Community of Practice: A peer-learning network of officials from 10+ Argentine cities, with equivalent coalitions launched in Brazil, Chile and Uruguay (2025–2026). The network produces publications, events and continuous exchange mechanisms to build lasting institutional capacity within local governments.

The horizontal model — where cities learn from each other, not from a central authority — was core to the design. Peer validation proved more effective at reducing adoption barriers than expert-driven top-down assistance.

This initiative began in August 2024, is currently active and expanding (mid-2026), with coalitions launched in Brazil, Chile and Uruguay in 2025–2026.

The Impact

1. Reach and scale. From a founding coalition of 10 Argentine cities in August 2024, the CIIAR model expanded to Brazil, Chile and Uruguay by 2026, creating the first multi-country Latin American alliance of local governments for AI. Founding Argentine cities include: Córdoba, Escobar, Mendoza, Neuquén, Pilar, Rosario, Salta, San Fernando del Valle de Catamarca, San Miguel de Tucumán and Tres de Febrero.
2. GovTech ecosystem. Over 350 Latin American startups applied to the GovTech Connect open call, demonstrating significant private-sector demand to engage with local government AI challenges. Demo Days were held across Argentina, Chile and Uruguay in 2025–2026.
3. Open resources. The free AI Maturity Diagnostic (Autodiagnóstico IA) is available to any local government in Latin America. The collective digital book "Artificial Intelligence in Local Governments" (RIL-CIIAR, 2025) documents practical, city-level lessons from member cities.
4. Policy development. CIIAR member cities in Argentina began advancing local AI governance ordinances in 2026, making them sub-national laboratories for responsible AI governance in a country without a national AI law.

Lessons Learned

What worked well:

• Horizontal peer learning outperformed expert-to-government models. Cities adopted AI faster when solutions were validated by peer cities facing similar constraints, rather than recommended by consultants or national bodies.
• The AI Maturity Diagnostic, as a first step: making a city's baseline visible before any talk of solutions, reduced political resistance and created a shared language. Cities stopped comparing themselves to large capitals and started building realistic roadmaps.
• GovTech Connect generated more genuine adoption than conventional public procurement — startups arrived with tested solutions, reducing the perceived risk for cautious public officials.
• Regional expansion was organic: cities in Chile, Uruguay and Brazil approached CIIAR on their own initiative after learning about the Argentine model.

What was harder than expected:

• Staff turnover and electoral cycles disrupted project continuity. Initiatives launched under one administration were sometimes orphaned after a change of government.
• Managing very different paces across cities with different digital maturity levels: advanced cities pushed for speed while less advanced cities needed foundational support.
• Moving from 'AI adoption' as political rhetoric to projects with clear, measurable outcomes remained the most persistent challenge.

What we would do differently:

• Define city-level impact metrics at project outset to build robust evidence beyond participation numbers.
• Formalise agreements between municipalities and govtechs from the Demo Day stage to ensure post-programme continuity.
• Engage technical teams within municipalities (not only senior officials) from day one to secure institutional ownership beyond individual champions.

CIIAR cities operate in a transitional, partially undefined regulatory environment. Three frameworks shaped the programme design:

Data protection: Argentina's Personal Data Protection Act (Law 25,326) sets requirements for processing citizen data, which directly influenced the criteria used to evaluate and select GovTech solutions in the programme.

Public procurement barriers: Argentina's procurement rules create structural barriers to acquiring innovative technology from startups. GovTech Connect was specifically designed to explore pilot and co-creation mechanisms that work alongside conventional procurement constraints.

Regulatory leadership from below: At CIIAR's launch (2024), Argentina had no national AI law. Rather than waiting for a national framework, CIIAR member cities began advancing local AI governance ordinances in 2026, becoming de facto policy laboratories for sub-national AI governance in Latin America.