NetworkDeployAI is an AI-powered decision support tool designed to optimize the deployment of next-generation telecom networks using existing pipeline sections (canalitzacions) managed by the Generalitat of Catalonia. The platform allows third-party telecom operators to repurpose these pre-dug pipelines, reducing construction time and costs while navigating regulatory hurdles. The tool integrates open-source pipeline data with supplementary information such as geospatial data (urban maps, zoning, population density), regulatory details (permit processes, policy guidelines), and technical data (network assets, vendor contracts). This data is used to create a dynamic digital twin of the existing infrastructure, simulating various deployment scenarios to assess pipeline capacity, network performance, and future scalability. Machine learning is employed to evaluate the suitability of pipeline sections for different network technologies (e.g., fiber-optic, 5G). Predictive models help forecast business impacts such as cost efficiency, ROI, and service resilience, while also recommending the best pipeline sections for network upgrades. A key feature of NetworkDeployAI is automated regulatory compliance. The platform decodes local regulations to generate tailored permit checklists, approval timelines, and vendor recommendations, making it easier for operators to comply with the approval process. It also analyzes past vendor data to propose cost-effective procurement strategies. The user-friendly web-based interface, allows operators to input project parameters and receive detailed blueprints with cost estimates, permit requirements, and performance predictions. By streamlining network design, regulatory compliance, and procurement, NetworkDeployAI helps improve the efficiency of network deployments and enhances connectivity for public services.
Category tags:"(Toni + Maria) Quite an interesting proposition to leverage underutilized assets - directly relevant to the hackathon tracks. More information on the specific ML methods used, algorithm and tuning would be nice to actually understand how the tool works and is intended to work. Does not fully use open source components (yes to data, but not to the rest of the tech stack). "
Maria Antonia Bravo
"Interesting use case but limited and NOT open source AI application (chatGPT). Either use AI to optimize network planning or use RAG to return answers more specific to network deployment and regulations in the country of interest"
Ivan Dotu
"Practical and innovative solution for optimizing telecom network deployment by leveraging existing infrastructure. An AI-powered digital twin is a timely and thoughtful solution. The solution's automated regulatory compliance and vendor recommendation features that streamline complex bureaucratic processes are also interesting. The presentation of the solution can be more polished but there is a lot of good substance to work with. Well done."
Anastasia Nedayvoda