How Luján de Cuyo is using AI to shift to proactive municipal maintenance services

A bald man with a red beard wearing a navy blue sweater gestures with both hands while speaking during a meeting around a wooden conference table. Four colleagues listen attentively, including a woman with glasses and blonde hair on the left, a woman with glasses and long dark hair in a light blue top, and a bald man in business casual attire on the right.

Summary

Luján de Cuyo, a city in northwest Argentina, is shifting from reactive public space maintenance to a proactive, data-driven model. Through its participation in the City Data Alliance (CDA), the city built the governance and data foundations needed to pilot an AI enabled detection service that identifies and georeferences potholes and road damage using cameras on municipal vehicles. Previously, maintenance depended on resident complaints and manual inspections, leading to delays, higher costs, and uneven service delivery. By using AI to detect risks early and guide repairs, Luján de Cuyo is improving safety, efficiency, and equity in a city service area residents experience every day.

 

Vision

Road conditions shape daily life in Luján de Cuyo, affecting safety, mobility, and trust in local government. For years, the city operated in a reactive cycle: waiting for complaints, dispatching inspectors, and only then scheduling repairs. This approach was costly and slow, and it left residents in peripheral neighborhoods waiting the longest for responses.

Mayor Esteban Allasino set a clear objective: identify risks in public space early and resolve them before they become accidents. In this vision, technology is not an end in itself, but a tool to deliver faster, fairer, and more transparent services. Residents should be able to see what issues are detected, how they are prioritized, and when they are addressed.

 

Approach

Luján de Cuyo’s AI-enabled maintenance work builds on institutional capacity developed through CDA, including a formal Data Committee, a citywide data inventory, and a strategic objectives monitoring system. With these foundations in place, the city focused on a single, visible problem with high resident impact: potholes and cracks in public roads. The approach centers on three elements:

  1. Early detection using existing assets: Municipal vehicles were transformed into mobile sensors equipped with cameras and GPS. As vehicles travel their regular routes, images are captured and securely processed. A computer vision model identifies road damage, classifies each issue by type and severity, and georeferences it on a map. This allows the city to detect problems systematically.
  2. Evidence-based prioritization of repairs: Detected issues feed directly into a maintenance backlog and operational dashboards. Crews use this information to prioritize interventions based on risk and severity, shifting from reactive fixes to preventive maintenance. This reduces long-term repair costs and improves service consistency across neighborhoods.
  3. Privacy, ethics, and equity by design: From the first prototype, the city embedded safeguards to build trust. Images are processed with automatic blurring of faces and license plates, raw image retention is minimized, and access is logged. Monitoring routes are designed to cover all neighborhoods, ensuring benefits reach pedestrians, cyclists, and people with reduced mobility, not only drivers or central areas. Aggregated data is published when appropriate to support transparency.

 

Impact

Although the AI-enabled detection service is currently in a pilot phase, early results and expected benefits are clear:

  • Faster response: Road issues that once took months to identify can now be detected within three to five days on pilot routes.
  • More preventive maintenance: The share of repairs completed before a resident complaint rose from approximately 10% to 40% on pilot routes.
  • Greater equity: Neighborhood-level monitoring helps ensure peripheral areas receive timely attention, not only those with higher complaint rates.
  • Improved efficiency: Risk-based prioritization reduces rework and directs resources to the most urgent issues.

Luján de Cuyo has committed to integrating this approach into regular maintenance operations, using data and dashboards to guide decisions and track performance over time.

 

Lessons

Luján de Cuyo’s experience shows that AI creates value when it strengthens everyday management, not when it operates as a black box. Starting with a visible, fixable problem helped build momentum and public value. Using existing municipal assets made rapid experimentation possible. Designing privacy, equity, and transparency from the outset reduced risk and increased trust. Most importantly, investing in internal capacity ensured the city could sustain and scale the work beyond a single pilot. Luján de Cuyo demonstrates how cities can use AI not as abstract innovation, but as a practical tool to deliver public services.