AI Readiness Index for public administration

The problem

Bulgarian public administration has strong potential for growth in Artificial Intelligence (AI), but the application of AI at the state administrative level has remained sporadic and uncoordinated, with a low level of knowledge, competence, confidence, and preparedness among civil servants. At the same time, there was no comprehensive picture of how public administration employees and institutions in Bulgaria perceive, use, and are prepared to implement AI.

The project addressed the need for an objective assessment of AI readiness in public administration, covering both the organisational environment and the individual capacity of employees. It also responded to the need to identify training needs, support requirements, and barriers to the effective and responsible use of AI in the public sector.

The solution

The Institute of Public Administration (IPA) developed and implemented the first AI Readiness Index for Bulgarian public administration — described as the first study of its kind in Bulgaria and the Balkans. The purpose was to provide a detailed and objective picture of the current state of AI readiness in public administration and to identify strengths, weaknesses, and areas for further development.

The Index was designed as a composite indicator combining two main dimensions: organisational readiness, reflecting policies, infrastructure, practices, and management attitudes towards AI; and individual capacity, reflecting employees' knowledge and skills in AI.

These two dimensions were measured through an online survey questionnaire. Responses were converted into numerical values, summarised into separate scores for organisational and individual readiness, and then combined into an overall index through a weighted average: 65% organisational readiness and 35% individual capacity. The overall index was scaled from 0 to 100 and interpreted through a four-level qualitative scale.

In addition to the readiness index itself, the study included a training needs module, which identified the most frequently cited skills gaps and areas for development, preferred forms of training (such as online platforms, webinars, face-to-face courses, and self-study), and the main barriers to participation in training, including lack of time, insufficient IT skills, and uncertainty about the practical application of knowledge.

The impact

1. First national baseline for AI readiness
The project produced the first national assessment of AI readiness in Bulgarian public administration, providing for the first time a detailed picture of attitudes, knowledge, practices, and institutional conditions related to AI across Bulgarian institutions.

2. Large-scale participation supporting representativeness
The survey included 6,254 respondents. The distribution of participants by key demographic characteristics showed close similarity to national administrative data, which supports the representativeness of the findings.

3. Clear demand for AI training identified
The training needs assessment demonstrated strong interest among public administration employees in further learning and development related to AI. The results highlighted demand for targeted capacity-building opportunities to support the effective use of AI in the public sector.

4. Methodology adopted internationally
The methodology generated strong external interest. It was adopted by Croatia and Romania, which conducted similar readiness index surveys among their own civil servants using the IPA methodology. There was also interest from non-governmental organisations in applying the methodology to other sectors such as education and health.

5. Existing capacity within the system confirmed
The results showed that there are already innovators and motivated experts within Bulgarian public administration, while also confirming that further efforts are needed in developing digital skills, modernising infrastructure, and creating internal rules and ethical standards for AI.

Lessons learned

• A readiness index can serve as both measurement and catalyst. The project created the first objective and structured tool for assessing AI readiness in Bulgarian public administration. Beyond producing data, the process itself had value: providing individual reports to participating institutions encouraged each one to examine its internal situation more closely and to undertake its own needs analysis in relation to AI readiness, skills, and organisational capacity. In this way, the index served not only as an assessment tool, but also as a starting point for institutional self-reflection and planning.
• The methodology travels well. The fact that Croatia and Romania subsequently adopted the methodology suggests that a well-designed readiness index can be replicated across countries with relatively limited adaptation. For other governments considering similar assessments, this demonstrates that investing in a sound methodology can produce returns beyond the originating country.
• Self-assessment has limits that need addressing. The methodology currently relies on self-assessment, which has known limitations. An important next step is to monitor over time how self-reported readiness corresponds to actual practice and institutional reality. Complementing survey data with real operational data would make the index more robust and provide a more accurate picture of AI readiness in public administration.
• Institutional capacity to maintain the methodology matters. Further progress depends on building on IPA staff's existing capacity by providing time and opportunities for deeper engagement with emerging methodologies for AI readiness measurement and with peer institutions working in this field.

The project was developed in a broader international, European, and national policy context related to artificial intelligence and digital transformation.

At the international and European level, the text refers to:

• the OECD toolkit for trustworthy AI
• the UNESCO framework for AI competencies
• The EU Artificial Intelligence Act, which sets requirements for the safe and ethical use of AI
• studies and principles from the European Commission’s Joint Research Centre.

At the national level, the project is linked to:

• Bulgaria’s broader digital government transformation efforts;
• the National Development Program Bulgaria 2030.