Name in original language
Asistente IPS Córdoba: agente de análisis de políticas de IA para la gobernanza del progreso social.
Initiative overview
The Province of Córdoba measures the Social Progress Index across all 427 of its localities over two census periods, producing one of the most granular subnational quality-of-life datasets in Latin America. Despite the richness of this data, a persistent and well-documented gap existed between data production and data use in public administration: mayors, municipal officials, and local planners often lacked the time, technical capacity, or analytical tools to translate 57 social indicators into concrete policy decisions. Complex statistical products were being generated but not acted upon. The IPS Córdoba Agent was developed specifically to close this gap — not by simplifying the data, but by building an intelligent analytical layer on top of it that is capable of reasoning across indicators, localities, time periods, and policy frameworks simultaneously.
The initiative has three core objectives. First, to democratise access to complex social intelligence: any local official — regardless of technical background — should be able to query their territory's social performance in natural language and receive a rigorous, grounded analysis in return. Second, to support evidence-based policy design at the local level by connecting social diagnoses directly to actionable intervention options drawn from provincial development frameworks. Third, to strengthen a culture of data-driven governance across Córdoba's municipalities, embedding AI-assisted analysis as a routine part of local decision-making rather than an exceptional resource.
The IPS Córdoba Assistant is designed and should be understood as a policy analysis agent, not a retrieval chatbot. The distinction is substantial and consequential. A conventional chatbot answers discrete questions by retrieving pre-existing text. This agent, by contrast, performs multi-step analytical reasoning: it interprets a locality's indicator-level results, identifies patterns of improvement or deterioration across dimensions and sub-dimensions, benchmarks performance against provincial averages, infers which population groups are most affected by specific social gaps, and generates policy recommendations grounded in a curated knowledge base that includes academic literature on human development, the provincial development model, and validated intervention frameworks. The agent synthesises all of this into a coherent, contextualised policy narrative — a form of analytical output that would otherwise require a trained policy analyst.
