Data-driven lifecycle governance for intelligent public lighting: a multi-city deployment study

Purpose This practitioner case study addresses the persistent gap between smart city aspirations and implementable urban operator practice by examining how artificial intelligence (AI)-enabled methods can be operationalised across the full infrastructure lifecycle within real infrastructure operator workflows. Design/methodology/approach The framework integrates four AI-enabled operational tools – GeoAI satellite asset mapping, a fuzzy-logic engine encoding national lighting standards, a Bayesian predictive maintenance model designed for data-scarce contexts and an adaptive lighting control system. The artefacts were deployed and evaluated through naturalistic pilots and retrospective analyses in municipalities in Italy, Spain and Chile. Findings Deployment evidence (operationally observed, not from controlled experiments) shows 82–92% asset detection accuracy across sites, ∼70% design effort reduction, ∼20% maintenance efficiency gain and ∼35% energy savings beyond light-emitting diode baselines. Cross-site deployment yielded five transferable design principles for scalable, regulation-aware governance of AI-enabled civic infrastructure. Practical implications The framework provides actionable guidance for municipal administrators and infrastructure operators pursuing AI-enabled infrastructure governance, contributing to SDG 7 (Affordable and Clean Energy) and SDG 11 (Sustainable Cities and Communities). Originality/value This study contributes an operationally validated governance framework that moves beyond isolated smart city pilots by embedding heterogeneous AI tools into systemic lifecycle management under real-world institutional and regulatory constraints.