The OECD.AI Policy Navigator

Our policy navigator is a living repository from more than 80 jurisdictions and organisations. Use the filters to browse initiatives and find what you are looking for.

Agent MIA (Argentine State's Artificial Intelligence Assistant)


Added by:   OECD analyst
Added on:   17 Jul 2026
Updated by:   OECD analyst
Updated on:   21 Jul 2026

MIA is integrated into the national government's chatbot (TINA). It was conceived as a support agent for citizen services management based on information security, administrative transparency, and human oversight principles. It offers personalized responses according to each user's needs and context. It is based on LLAMA 4 (Meta) for speed, low operating cost, and open-source nature. Currently in its initial phase with two use cases: SMEs and Mi Argentina (citizens' platform).

Name in original language

MIA (Asistente de Inteligencia Artificial del Estado Argentino)

Initiative overview

The initiative seeks to address inefficiencies in public services by modernizing them to achieve a more transparent, efficient, and citizen-centered government. Specifically, it addresses the following points:

  •  It aims to reduce waiting times and simplify digital processes for citizens.
  •  It solves the problem of misinformation by drawing on a closed knowledge base that uses only official data, thus ensuring accuracy and avoiding bias.
  • Unlike traditional chatbots that only respond, MIA can act: query data, generate reports, and perform administrative tasks under human supervision.

MIA's main objectives focus on improving the interaction between the government and society: 

  • Supporting SMEs and assisting the 26 million users of the Mi Argentina app with their digital transactions.
  • Offer a more agile, efficient, simple, and secure conversational experience.
  • Save time for both users and government support teams.

MIA is expected to evolve from a consultation assistant to a comprehensive government management tool: • We are working to expand the initiative and achieve integration into critical environments such as the Electronic Document Management System (GDE), Online Procedures (TAD), and the official portal Argentina.gob.ar. MIA has been designed to scale to a wide range of topics and services. In the next 12 to 18 months, the goal is for these AI agents to significantly simplify people's lives and make public administration more efficient at all levels. 

Currently, it already registers more than 1.5 million conversations per month. Impacts are measured using internal performance metrics for continuous improvement and the exploratory application of UNESCO's RAM Methodology. In the future, MIA will evolve from conversational to agentic, executing administrative tasks and generating reports in systems such as GDE and TAD. Over the next 12-18 months, it is expected to be the cornerstone for positioning Argentina as a regional AI hub, enabling a government that not only responds but also acts to simplify citizens' lives and optimize public administration.

Other relevant details

Implementation has faced critical barriers, ie: • Lack of technical capabilities as a limitation for technology adoption • Need for quality data and interoperable systems • Lack of trust and difficulty adapting to change in public organizations • Concern that over-regulation could kill innovation Lessons learned from implementation: • Public-private collaboration is essential to improve quality and reduce implementation times. • Technology should solve specific citizen needs. • Interoperability as a foundation. • Sometimes, existing and scalable solutions generate more impact than immature innovations.