Public administration is on the verge of a new era, shaped by the rapid rise of artificial intelligence (AI). On a global scale, governments  are already taking advantage of AI's potential for better services and governance, while embracing the principles of responsible and humancentered use of these technologies. The OECD's toolkit for trustworthy AI and initiatives such as the UNESCO framework for AI competencies outline a vision for skills development and policies, while the EU Act on Artificial Intelligence sets out specific requirements for safe and ethical use.

Artificial Intelligence Readiness Index of the State Administration, the first-of-its-kind study in Bulgaria and the Balkans. IPA developed the methodology for the Index and conducted the first study in 2025 in wich 6,254 civil servants participated.

The AI Readiness Index is a composite indicator developed to assess the overall capacity of administrations to implement artificial intelligence. It combines two main aspects of readiness:

The two components were measured using a questionnaire (online survey) containing groups of questions on the relevant topics.

The index's methodology has generated considerable interest among IPA-related institutions in the Balkans. As a result, Croatia and Romania have conducted similar surveys of public administration employees, based on the methodology provided by IPA. Interest has also come from non-governmental organizations to apply the methodology to other sectors like education, health etc.

The second national Artificial Intelligence Readiness Index study tracks changes in attitudes, knowledge, and practices related to the use of AI in Bulgaria’s public administration between 2025 and 2026. The data were collected from 5,475 employees at central, regional, and municipal levels.

In 2026, the overall Index score increased, while the administration remained in the “capacity-building” stage. The most notable progress was observed in the practical use of AI and employees’ readiness to work with AI tools.

At the same time, the institutional environment is developing more slowly. Limited access to approved tools, the lack of clear internal rules, and insufficient knowledge sharing continue to be among the main barriers.

The findings show that the sustainable adoption of AI in public administration requires not only stronger skills and more training, but also clear responsibilities, secure access to AI tools, and mechanisms for institutional learning.

The raw data from the study are available in the open research repository Zenodo, DOI: 10.5281/zenodo.22026174.

National Study AI Readiness Index 2026

In 2025, the Institute of Public Administration conducted the first AI Readiness Index study. The overall index for the administration is 45 of maximum 100, placing it in the "building capacity" category. On this basis, the foundations for the effective use of AI can be considered to have been laid, but further action is needed: targeted investment in training, resource support, and strategic leadership to accelerate the uptake of AI technologies in government.

The results are both a reason for realism and optimism. The findings clearly show that the foundations have already been laid - there are innovators and motivated experts in the system. At the same time, areas emerged where targeted efforts are needed: developing digital skills, modernizing infrastructure, and creating internal rules and ethical standards for AI.

The AI Readiness Index is a valuable tool with which we can measure our progress in the coming years and compare our development with the best global practices.

AI Readiness Index Report 2025


The translation of this document was made by AI enabled technologies - DeepL Translator. and reviewd by experts of IPA.

чИПА