Georgia’s National Statistics Office (Geostat) has begun deploying artificial intelligence across its official statistics work, the agency announced on July 20, 2026.

A chatbot for public queries

The centerpiece for ordinary users is a new AI-powered chatbot on Geostat’s website. It answers questions about economic, social, and demographic indicators in natural language, pulling relevant figures from the agency’s publications and databases rather than requiring visitors to dig through spreadsheets themselves.

Behind the scenes

Geostat says AI is also now embedded in internal operations: reviewing and updating methodological documentation, classifying administrative data, monitoring databases, and supporting statistical analysis. On the technical side, the agency’s IT team is using AI for software development, code generation, algorithm design, and system testing.

A classifier that understands plain descriptions

A related tool, launched separately on July 14, lets businesses and officials find the right NACE Rev.2 code — the EU-derived system used to classify economic activities — by typing or speaking a plain description of what a company does. The AI assistant asks follow-up questions when the input is ambiguous, and the portal runs in both Georgian and English.

Part of a wider digital push

“Artificial intelligence has already become an important tool in the development of official statistics,” Geostat’s executive director, Gogita Todradze, said in the announcement. “Our objective is to further improve the efficiency of statistical processes through the application of modern technologies” while delivering “fast, innovative, and high-quality services.”

The move comes as AI adoption spreads across Georgia’s public and private sectors: a recent BAG survey found most Georgian companies are already adopting AI in some form, part of a broader build-out of the country’s AI startup and government ecosystem. Geostat’s rollout adds a concrete government case study to that trend — a statistics office using AI not for headline-grabbing models, but for the routine, unglamorous work of classifying data and answering citizen queries.

Geostat did not disclose which AI vendor or models power the chatbot or the NACE portal, nor did it specify a budget for the initiative.