Adoption of artificial intelligence and digitalisation in Bulgaria: a 2026 analysis

Bulgaria is not behind on everything. The country has good connectivity, national digital systems that work, a strong research core in INSAIT, the forthcoming BRAIN++ AI factory and product companies selling on international markets. The core problem is not an absence of technological assets but their limited diffusion — into mainstream business, into public administration beyond a few flagship systems, into most universities and into the regions. (Sources: EC, Digital Decade 2026; EuroHPC, BRAIN++; World Bank)

On the key indicators the country remains well below the EU average. In 2025, 38% of people aged 16–74 had at least basic digital skills, against 60% across the EU. Artificial intelligence was used by 8.55% of enterprises, against 19.95% in the EU. Online public services were used by 36% of citizens against 72% in the EU, and electronic identification by 12% against 52%. Research and development spending in 2024 stood at 0.77% of GDP against 2.24% for the EU. (Sources: Eurostat, digitalisation 2026; Eurostat, AI in enterprises; Eurostat, e-government; Eurostat, R&D)

These gaps point to a systemic deficit in scaling. Bulgaria produces isolated "islands of excellence" but does not yet convert them quickly enough into higher productivity, better public services and stronger technology transfer. Start-up activity and capital are concentrated in Sofia, and the links between universities, public data and industry remain uneven. (Sources: World Bank; EC, Digital Decade 2026)

For 2026–2030 the priority has to shift from individual projects to broad adoption. That means support for digitalisation and AI in small and medium-sized enterprises, access to quality data and testing environments, more pragmatic transfer from universities to business, and a clear framework for trust and compliance with European rules.

The AI Council at the Bulgarian Chamber of Commerce and Industry can serve as a practical coordination platform between business, universities and the public sector. The greatest added value lies in concrete sectoral pilots, shared outcome indicators, training for management teams, and proposals for a more predictable regulatory and institutional environment.

Where Bulgaria stands today

The European Commission has not published DESI as a standalone composite index since 2022. From 2023 the monitoring has been folded into the Digital Decade framework, so any current comparison has to be made indicator by indicator rather than through a single overall score. In the last full DESI ranking, for 2022, Bulgaria placed 26th of 27 member states. The 2026 country report continues to record weaknesses in digital skills, technology adoption in SMEs, innovation and cybersecurity, alongside strong connectivity. (Sources: EC, DESI; DESI 2022 — Bulgaria; EC, Digital Decade 2026)

A comparative profile on the key indicators

Denmark and Finland are used as reference points because they are small open economies with high digital maturity, strong AI results, widespread use of electronic services and high skill levels. The figures below are the latest available as of July 2026; the reference year is given for each indicator.

Note: "n/a" means not applicable. DESI is a composite index only up to 2022. For Government AI Readiness there is no official EU average, and the index methodology differs from that of DESI. (Sources: EC, DESI; Eurostat, digitalisation 2026; Eurostat, AI; Eurostat, e-government; Eurostat, R&D; Oxford Insights 2024)

What matters is not any single indicator but the overall pattern. The distance to Denmark and Finland is systemic: roughly four to five times on enterprise use of AI, close to three times on online public services and about eight times on electronic identification. That reflects differences not only in technology but in service design, management capacity, skills and trust.

A closer reading of the indicators

Basic digital skills are improving, but too slowly. Bulgaria rises from about 31% in 2021 to 38% in 2025, while the EU moves from 54% to 60%. The gap is not closing fast enough. With generative AI spreading widely, this is a direct obstacle to adoption in SMEs and in the administration, because missing basic skills raise the risk of errors, low returns and dependence on outside suppliers. (Sources: DESI 2022 — Bulgaria; Eurostat, digitalisation 2026)

In SMEs the deficit is more practical still. In 2025 only 38% of Bulgarian SMEs reached at least a basic level of digital intensity, against 71% in the EU, 92% in Denmark and 94% in Finland. A firm without dependable core systems — for resources, customers, documents, data and cybersecurity — will struggle to turn AI into a productive tool. In that context AI often stays a demonstration or a standalone subscription rather than a change to the process. (Source: Eurostat, digitalisation 2026)

The gap in research and innovation is structural. In 2024 Bulgaria spent 0.77% of GDP on R&D against 2.24% for the EU, 3.01% for Denmark and 3.22% for Finland. A further signal is the size of government budget allocations for R&D: €38.3 per person in Bulgaria against an EU average of €284.7 in 2024. That constrains modern infrastructure, the ability to attract and retain talent, and the number of applied projects with industry. (Sources: Eurostat, R&D; Eurostat, government budget allocations for R&D per capita)

On information and communication technology specialists the picture is mixed. Their share of employment in the EU reached 5.0% in 2025, and in Finland it is around 7.8%. Bulgaria has a competitive advantage in the higher share of women among ICT specialists — 25% in 2025. That does not solve the shortage of staff, but it offers a better base for widening the talent pool through training, reskilling and retention policies. (Source: Eurostat, ICT specialists)

The start-up ecosystem is more developed than the general innovation indicators suggest, but it is heavily concentrated. The latest World Bank assessment records about 348 active start-ups in Sofia — roughly 87% of the national total — and about €500 million in assets managed by local venture capital funds at the end of 2024. At the same time venture investment in 2024 fell to about €50–55 million, and under 10% reaches Series B and later stages. Bulgaria can create strong companies, but it does not yet provide a reliable capital bridge to international scale-up. (Source: World Bank)

In the public sector, the existence of digital systems and their actual use need to be kept apart. Bulgaria has digitised key processes, including public procurement and a substantial part of health information, yet in 2025 only 36% of citizens used a public authority's website or application and 12% used electronic identification to access services. The next stage is therefore not simply building more systems but better integration, fewer administrative steps and a clearer benefit to the user. (Sources: Eurostat, e-government; EC, Digital Decade 2026)

AI adoption in business is growing, but from a low base. The share of enterprises using at least one AI technology rose from about 6.5% in 2024 to 8.55% in 2025. Over the same period the EU reached 19.95%, Denmark 42.03% and Finland 37.82%. Bulgaria is not one year behind; the difference is one of scale, and of how far firms can turn data and automation into everyday processes. (Source: Eurostat, AI in enterprises)

Policies and flagship projects

Bulgaria has a broad strategic framework. The Concept for the Development of Artificial Intelligence to 2030 sets the overall vision; the updated "Digital Transformation of Bulgaria 2024–2030" links national goals to the European Digital Decade; the Innovation Strategy for Smart Specialisation 2021–2027 directs research and innovation priorities; the National SME Strategy has been extended to 2030. The problem is not a shortage of documents but weak coordination between them, limited prioritisation and uneven implementation. (Sources: AI Concept to 2030; Digital Transformation 2024–2030; ISSS 2021–2027; SME Strategy 2021–2030)

The national Digital Decade roadmap contains 60 measures with a total budget of €2.19 billion. The European Commission notes that 48% of the measures expire by the end of 2026; that amounts to €597 million of public funding, or 27% of the roadmap's public budget. The risk is a gap between the measures now ending and the next programming cycle. What is needed is a mid-term review, a public status for the key measures, and reallocation of resources in advance towards activities with a proven effect. (Source: EC, Digital Decade 2026)

The regulatory framework is also entering a practical phase. The prohibited practices and the AI literacy requirements have applied since February 2025, the governance rules and those for general-purpose models since August 2025, and the main body of the AI Act applies from 2 August 2026. Later deadlines apply to some high-risk systems. Bulgarian organisations need clear national guidance, model contracts and procurement templates, sectoral testing environments and accessible compliance support — without unnecessarily stalling innovation. (Source: EC, AI Act)

The flagship projects that change the picture

BRAIN++ is the most significant new infrastructure asset. Bulgaria's AI factory will bring together the AI-optimised Discoverer++ supercomputer and a service centre for state institutions, educational organisations, researchers and companies. The project is worth around €90 million. Managed well, BRAIN++ can become a national accelerator for prototypes, Bulgarian language models, robotics, healthcare, agriculture and tools for trustworthy AI. Its economic effect will depend on transparent access rules, measurable sectoral results and active SME participation. (Sources: EuroHPC, BRAIN++; Ministry of Innovation and Growth, €90 million project)

INSAIT is already a source of internationally visible research results, talent and potential new companies. In 2026 the institute reports 17 papers accepted to the main programme of CVPR, first place in Eastern Europe and a place among the top four institutions in the EU on that measure. The national task is to connect that research capacity systematically to industrial doctorates, joint laboratories, proof of concept and university spin-offs. (Sources: INSAIT, CVPR 2026; World Bank)

Discoverer+ is the third key asset. Upgrading it with AI systems widens the scope for specialised services to business, science and the public sector. The supercomputer and BRAIN++ will have a meaningful economic effect only if they are connected to universities, European digital innovation hubs, industry clusters and concrete manufacturing and public applications. (Source: Ministry of Innovation and Growth, Discoverer+)

Cases from business, the public sector and universities

Business

Shelly Group shows how digitalisation can become a product in its own right rather than only an internal process. The company develops solutions for the internet of things, smart buildings, automation and energy management, connected through mobile and cloud applications. According to the World Bank's assessment, in 2025 the company's market value passed one billion US dollars. The lesson is that Bulgaria can build globally competitive companies when an engineering base is combined with an own product and a clearly solved customer problem. (Sources: Shelly, investor report 2025; World Bank)

Payhawk represents a different trajectory — a fintech platform that embeds AI in financial processes. In 2025 the company developed its "AI Office of the CFO" and AI agents for procurement, travel, expenses and payments. The World Bank notes that Payhawk reached unicorn status in 2022. What matters about the case is the existence of a Bulgarian enterprise AI product with an international market, not the company's valuation alone. (Sources: Payhawk, Fall 2025; World Bank)

Postbank is an example of staged adoption in a traditional sector. The bank uses EVA as a digital assistant in Bulgarian and English and applies AI tools in recruitment. The approach is instructive: organisations often see the fastest results when they start with customer service, human resources and knowledge work before moving to riskier core systems. (Source: Postbank, EVA)

Public sector

CAIS EPP, the centralised electronic public procurement system, is a mature example of process digitalisation at national scale. In 2025 the system carried 111,415 notices and received 111,355 tenders. At the same time the share of procedures with a single bid was 39%, against a target of 23% under the National Recovery and Resilience Plan. This marks an important boundary: digitising a process improves traceability and efficiency, but it does not automatically fix problems of competition and market quality. (Source: Public Procurement Agency, annual report 2025)

The National Health Information System is the second strong public case. As of November 2025 it had processed over 615 million electronic health records, serves close to 3 million requests a day, more than 21,000 doctors actively use the eRx application, and active citizen users of eZdrave number around 240,000. This is a genuine state platform at scale, and it can support a next wave of trustworthy AI — given clear rules on data quality, access, security and human oversight. (Source: Information Services, NHIS)

Electronic justice and the Unified Information System of the courts are the third example. On November 2025 figures, over 2.4 million electronic cases have been filed and heard over five years, more than 8.7 million electronic acts issued, the system is used by over 12,300 users and connected to more than 15 external systems. The most suitable first applications of AI here are speech-to-text, document workflow support, search and analytical assistance — not opaque automation of judicial decisions. (Source: Information Services, electronic justice)

Universities and research transfer

In Bulgaria the strongest academic capacity in AI is concentrated in a small number of structures, whereas the leading European systems have broader networks of universities, industrial partnerships and sustained funding for transfer. Low national R&D intensity and limited commercialisation mechanisms mean that research successes spread only with difficulty beyond individual centres. (Sources: World Bank; Eurostat, R&D)

Sofia University concentrates the most visible academic core through INSAIT and GATE. GATE was established as an autonomous structure for big data and smart society applications, while INSAIT produces internationally visible publications and early commercialisation results. This model shows how a university can be at once an educational institution, a research infrastructure and a basis for creating new companies. (Sources: GATE; INSAIT, CVPR 2026; World Bank)

The Technical University of Sofia builds the engineering and applied base. The ICARAI 2025 international conference, intensive AI programmes and projects in robotics, communications and signal processing matter for the workforce that manufacturing, embedded systems and industrial automation need. That capacity should be linked to more joint laboratories and to projects commissioned by enterprises. (Source: Technical University of Sofia, ICARAI 2025)

The Medical University of Pleven illustrates the scope for AI in medical education and clinical practice. In 2025 the university introduced an AI application in its library and developed activities in robotic surgery and telemedicine. Healthcare is one of the most promising fields for trustworthy AI in Bulgaria, but it requires clinical expertise, quality data, validation and ethical rules together. (Source: Medical University of Pleven)

Barriers and opportunities

A system of interlocking barriers

The first barrier is institutional and regulatory predictability. The core problem is not a formal absence of rules but their uneven application, slow commercial disputes, difficult insolvency, weak protection of intellectual property and unclear rules on ownership of results from publicly funded research. In AI, legal uncertainty raises transaction costs and discourages investors, universities and scaling companies. (Source: World Bank)

The second barrier is skills. The shortage is not confined to programmers. What is needed are managers, engineers, data specialists, applied researchers, teachers and civil servants who can turn the technology into a working process. Demographic decline, outward migration of qualified people and a weak match between education and demand turn AI policy into policy for the general capacity of the economy. (Sources: World Bank; Eurostat, digitalisation 2026)

The third barrier is funding beyond the early stage. Bulgaria has funds and programmes to start companies, but under 10% of venture investment reaches Series B and later stages. A capital and market bridge is needed between a proven product and international scale-up. Without it, the strongest companies move key functions abroad precisely when they begin to create the most value. (Source: World Bank)

The fourth barrier is access to data and how little of it is put to use. Bulgaria has significant systems such as the NHIS, RegiX and CAIS EPP. RegiX connects dozens of registers and provides a basis for exchange between authorities, but that does not automatically mean the data are well documented, standardised and fit for analysis and machine learning. Common rules on data governance, application programming interfaces, anonymisation, synthetic data and secure testing environments are required. (Sources: State e-Government Agency, RegiX; Information Services, NHIS; Public Procurement Agency, annual report 2025)

The fifth barrier is regional concentration. Growth, start-up activity and much of the specialised talent sit in Sofia and, to a lesser degree, Plovdiv. Outside those centres, infrastructure, management capacity and the links between universities and firms are uneven. Bulgaria has strong digital centres but not yet a national network dense enough to diffuse technology. (Source: World Bank)

The sixth barrier is trust. Low use of electronic identification and of public electronic services shows that a technological supply does not automatically become use by citizens and business. In AI, trust requires clear accountability, human oversight, traceability, explainability and the possibility of challenge. Without those elements, public deployment will generate resistance rather than value. (Sources: Eurostat, e-government; EC, AI Act)

The opportunities are real

Bulgaria has a rare combination for a country of its size: a high-calibre research core, European infrastructure through BRAIN++, working national digital systems holding large volumes of data, product companies with international markets and a comparatively high share of women in ICT professions. That makes the next phase of development attainable — but only if these assets are used in a coordinated way. (Sources: INSAIT, CVPR 2026; EuroHPC, BRAIN++; Eurostat, ICT specialists; World Bank)

The right reference point is not a mechanical copy of large economies but the model of small, coordinated European states. Three practical lessons can be drawn from Denmark and Finland: a strong digital core in the public sector, a standing connection between science and enterprises, and a consistent focus on mass adoption in SMEs. The difference is made by execution at scale, not by the number of strategic documents. (Sources: Eurostat, digitalisation 2026; Eurostat, e-government)

What to do now

The recommendations are ordered by strategic importance. For planning purposes an indicative scale of the public and blended resource required is used: low — up to about 10 million leva; medium — about 10–50 million leva; high — over 50 million leva. What matters is that the investment builds mechanisms for diffusing capacity rather than a series of disconnected pilot projects.

Priority recommendations and timeline

P1. A national programme for AI adoption in SMEs — through vouchers, sectoral consultancy and the European digital innovation hubs. Every project should cover data, process and cybersecurity, not merely a licence for a tool.
Timing: start in 6–12 months, scale over 36 months. Lead actors: Ministry of Innovation and Growth, Ministry of e-Government, BCCI, EDIHs, branch organisations, municipalities. Resource: medium to high.

P2. A national coordination unit for AI governance and procurement — procurement and contract templates, risk classification, model cards, human oversight, audit trails and readiness for the AI Act.
Timing: 6–12 months. Lead actors: Ministry of e-Government, Council of Ministers, Commission for Personal Data Protection, Public Procurement Agency, sectoral regulators, BCCI. Resource: low to medium.

P3. Sectoral data spaces and testing environments for healthcare, justice, energy and industry, including standards for anonymisation, synthetic data and access control.
Timing: 12–36 months. Lead actors: Ministry of e-Government, Ministry of Health, Ministry of Justice, Ministry of Energy, National Statistical Institute, universities, Sofia Tech Park. Resource: high.

P4. Unified rules for university intellectual property and a national fund for proof of concept and transfer from universities and the Bulgarian Academy of Sciences to enterprises.
Timing: 12–24 months. Lead actors: Ministry of Education and Science, Ministry of Innovation and Growth, Fund of Funds, universities, Bulgarian Academy of Sciences. Resource: medium.

P5. A scale-up financing instrument for Series A and B rounds, and use of the public sector as a first customer for Bulgarian deep-tech and AI companies.
Timing: 12–36 months. Lead actors: Ministry of Innovation and Growth, Fund of Funds, Bulgarian Development Bank, EIF/EIB, Public Procurement Agency. Resource: high.

P6. A large-scale AI literacy programme and practical upskilling for managers, SME employees, civil servants, teachers and university lecturers.
Timing: start within 6 months, then continuous. Lead actors: Ministry of Education and Science, Ministry of Labour and Social Policy, Institute of Public Administration, BCCI, universities, employers' organisations. Resource: medium to high.

P7. A regulatory sandbox and a compliance advisory centre for high-risk AI, e-government, health technology and financial technology.
Timing: 6–18 months. Lead actors: Ministry of e-Government, sectoral regulators, Commission for Personal Data Protection, Bulgarian National Bank, Financial Supervision Commission, BCCI. Resource: low to medium.

P8. Five national missions with a measurable outcome: AI in healthcare, justice, municipal services, energy efficiency and industrial automation.
Timing: 12–48 months. Lead actors: Council of Ministers, line ministries, municipalities, universities, business. Resource: high.

The first package for immediate action should cover P1, P2 and P4. P1 addresses the central deficit — the weak diffusion of technology into mainstream business. P2 creates a predictable framework, without which the public sector will oscillate between excessive caution and disconnected pilots. P4 turns research results into products, licences and new companies. The three measures need to run in parallel; otherwise BRAIN++, INSAIT and individual company successes will not add up to broad economic change. (Sources: EC, Digital Decade 2026; World Bank; EuroHPC, BRAIN++)

The roles of the key actors

No single actor can deliver the transformation alone. The state sets the rules, manages public data and shapes a significant part of demand. Universities and research organisations supply talent, research and infrastructure. Business carries the responsibility for adoption, productivity and international markets. BCCI and the branch organisations can connect these three systems and turn shared goals into sectoral programmes.

For the AI Council at BCCI the most useful role is to coordinate the diffusion of AI across sectors, skills and adoption rules. That requires standing working groups, regular meetings between firms, universities and the administration, and a common set of key performance indicators:

• The share of participating SMEs that adopt a solution and are still using it after 12 months.
• The measured change in productivity, processing time, quality or the cost of the process.
• The number of accessible datasets and testing environments with clear rules of use.
• The number of university proofs of concept, licences and new companies.
• The number of managers and employees trained, and the share of organisations with internal rules for responsible use of AI.

Immediate next steps for the Council

1. Within three months, set up working groups on SMEs, skills, university transfer and regulatory questions, each with a named lead and a timetable.
2. Conduct structured interviews with representatives of leading SMEs, universities, municipalities and public institutions, in order to select problems with clear economic or social value.
3. By month six, start at least two sectoral pilots with predefined indicators, a budget, a process owner and a scaling plan.
4. By the end of 2026, publish a short scoreboard report on progress, results, lessons learned and the changes needed in regulation, funding and public procurement.

Conclusion

Bulgaria has a real opportunity to become a regional centre for applied AI in South-East Europe. The assets are there: INSAIT, BRAIN++, the expanded supercomputing infrastructure, national systems in healthcare, public procurement and justice, and product companies with international success. The weakness is that these assets are not yet converting quickly enough into mass adoption in SMEs, in the regions and in public services. (Sources: INSAIT, CVPR 2026; EuroHPC, BRAIN++; EC, Digital Decade 2026; World Bank)

The goal should therefore not be for Bulgaria to have a handful of impressive projects. The goal is for AI and digitalisation to raise the productivity of thousands of enterprises, the quality of dozens of public services and the economic value of university research. It is precisely in connecting business, the state and the universities that the AI Council at BCCI can have the largest and most measurable impact.