The Statistical Office of the Republic of Serbia, in cooperation with Microsoft, successfully hosted the AI-Ready Dissemination Project Sprint from April 20 to 22 at the Microsoft Development Center in Belgrade. The event brought together the global official statistics community to jointly explore and advance approaches to AI-driven dissemination.
The sprint gathered over 30 in-person participants from 12 countries, along with more than 50 additional participants joining online from across the world. Attendees included representatives of national statistical offices, international organizations, and technology partners, with active participation from Microsoft experts contributing to both technical discussions and strategic perspectives.
Over the course of three days, the program combined expert presentations, working group sessions, and plenary discussions. Key topics included defining quality dimensions for AI-generated statistical content, establishing common approaches for monitoring AI performance, and improving data and metadata structures to make official statistics more accessible for AI systems.
Participants also explored AI search architectures and dissemination chatbots, including natural language access to statistical data through intelligent platforms. Important discussions addressed interoperability frameworks such as SDMX, emerging concepts like Model Context Protocol (MCP) and agent-based data access, as well as the foundations of a reference architecture for AI-ready dissemination systems.
In addition, the sprint tackled broader strategic themes, including AI sovereignty, governance, and trustworthiness, with a strong emphasis on ensuring that AI solutions remain aligned with the core principles of official statistics - particularly accuracy, transparency, comparability, and confidentiality protection.
The event concluded with agreement on key next steps, including further development of the project report, preparation of an international industry roundtable, and continued collaboration among participating institutions throughout the year.
By hosting this sprint, SORS reaffirmed its role as an active contributor to international initiatives and its commitment to developing innovative, trustworthy, and globally aligned approaches to statistical dissemination in the AI era.
Gallery
