Romanian firm Flashnet has launched software to simplify management of large-scale smart city infrastructures through natural language interaction. The AI Module is intended to help municipal operators manage complex data streams from street lighting and IoT sensors.

The tool employs an isolated Large Language Model (LLM) trained on proprietary data sets, allowing operators to filter devices and generate reports using plain-language queries. The system is designed to reduce the technical learning curve for Control Management System (CMS) users.

The Flashnet AI Module includes natural language data search and device filtering, heuristic anomaly detection to identify unexpected system errors, and predictive maintenance insights based on historical data patterns. It provides an open API architecture for integration with multi-vendor platforms, and support for infrastructures with millions of connected devices.

Nicola Crespi, Flashnet’s Chief Product Officer, said the AI is designed to recognise data patterns and predict system failures faster than manual monitoring. The module is engineered to operate across different vendor platforms to support neutral smart city operations.

The application is built for enterprise-grade security and low-latency performance. Flashnet stated that the AI layer can be integrated into existing urban management applications to provide a unified intelligence interface for diverse IoT sensors.