AGORA-AI (Assistant for Grounded Operational Recommendations in Urban Administration) is an innovative AI-powered decision-support system designed to help municipalities and public authorities address complex urban challenges through scientifically grounded recommendations and alternative scenario simulations.

The project is developed by DOTSOFT S.A. and builds upon the NOON-IoT smart-city digital twin platform, which collects and analyses operational data from IoT sensors and municipal information systems. AGORA-AI extends the platform’s capabilities by combining real-world urban operational data with scientific knowledge through Knowledge Graphs, Large Language Models (LLMs), and Retrieval-Augmented Generation (RAG) techniques.

Through a natural-language interface, municipal administrators can describe urban challenges, such as traffic congestion or increased energy consumption in public buildings, and receive ranked recommendations for potential interventions. These recommendations are based on comparable cases from other cities and are accompanied by projected impact estimates, explicit uncertainty levels, and verifiable scientific references.

A key feature of AGORA-AI is its what-if scenario simulation capability, which enables public authorities to evaluate alternative solutions before implementation. The initial expansion of the system focuses on urban mobility and environmental management, with the potential for future applications in additional areas, such as energy, waste management, and water supply networks.

The project includes the planned integration of European AI models and vector search services through the AI-on-Demand (AIoD) platform, the enhancement of multilingual capabilities, and the evaluation of the system in a real-world municipal environment. The objective is to advance the existing operational prototype from Technology Readiness Level (TRL) 7 to TRL 8, with a strong emphasis on transparency, interoperability, and European digital sovereignty.

Type and Scope of Work to Be Provided

The planned contribution of DOTSOFT S.A. focuses on the development, technological integration, and pilot evaluation of AGORA-AI as a comprehensive decision-support solution for public administration. More specifically, it includes:

• Knowledge Graph Expansion: Enriching the existing knowledge graph with urban challenges, proposed interventions, city characteristics, and measurable outcomes, leveraging scientific publications and operational data from the NOON-IoT platform.
• European AI Integration: Integrating the system with European Large Language Models and embedding/vector search services through the AIoD platform, replacing existing dependencies on non-European models.
• Similarity Analysis and Scenario Simulation: Enhancing semantic and structural similarity mechanisms to identify comparable urban cases and estimate the expected impact of different interventions, while explicitly representing uncertainty.
• Multilingual Support: Extending the natural-language interface to support six European languages: Greek, English, French, Spanish, Polish, and Portuguese.
• Interoperability and Data Integration: Connecting AGORA-AI with the NOON-IoT platform and other compatible smart-city infrastructures through open standards and technologies such as NGSI-LD and FIWARE.
• Pilot Deployment and Evaluation: Deploying and testing the system in a real-world municipal environment, in collaboration with a public-sector organisation (AI Adopter), while evaluating recommendation quality, user confidence, and the effectiveness of decision support.
• Transparency and Responsible AI: Implementing mechanisms for recommendation traceability, data protection, and human oversight, ensuring that final decisions remain under the authority of responsible public officials.

Through AGORA-AI, DOTSOFT aims to enhance the efficiency, transparency, and sustainability of urban administration by transforming smart-city data into evidence-based, actionable recommendations. The project seeks to support the transition from simply monitoring urban challenges to addressing them proactively and responsibly, leveraging trustworthy European Artificial Intelligence technologies.