Atos announced it was awarded by the Madrid City Council the maintenance, the evolution and support services of its Artificial Intelligence (AI) platform, within the framework of the Madrid Artificial Intelligence Initiative (MAIA). The goal of the project is to optimize administrative processes, improve municipal services and internal management, and offer citizens more efficient, proactive and personalized value-added services. The generative AI contemplated in the contract will allow the development of an ecosystem of artificial intelligence agents capable of helping citizens to solve their needs from the Madrid City Council.
The contract, with a duration of three years and the possibility of extension for an additional year, includes the corrective, adaptive and evolutionary maintenance of the different services that operate on the AI platform of the Madrid City Council, as well as operational support to users and the Computer and Communications Service of the Madrid City Council (Easydro). In close collaboration with the Madrid City Council, Atos will promote excellence and collaboration in the use of AI through technological surveillance, an innovation laboratory, a knowledge network, specialized trainings and technical and methodological support for the design and monitoring of artificial intelligence projects. In addition, Atos will develop new use cases and enhancements to the municipal AI platform, while resolving incidents and adapting AI systems to technical or regulatory changes, offering support for both day-to-day operations and end-users.
Among the many use cases that Atos is developing, examples such as Clear Communication and Pipeline of Documentary objects (text, audio, video) stand out, which highlight the potential of AI in public management. Clear Communication aims to improve the writing and understanding of administrative texts, ensuring that they are clear, accessible and understandable for citizens, in accordance with the Madrid City Council's Clear Communication Guide. Thanks to advanced language models, the system can analyze the context, suggest formulations and justify the proposed changes, thus strengthening transparency, efficiency and linguistic equity.
On the other hand, Pipeline of Documentary Objects allows the automatic processing of audio, video and text files, transforming their content into structured information through transcription and OCR, and generating vector representations (embeddings) that facilitate semantic searches and contextual queries. This capability accelerates the localization of relevant information and enables the integration of surge recovery (Retrieval Augmented Generation - RAG) systems to deliver accurate and contextualized responses.
