Who’s Actually Building the Adria Region’s AI Economy?
From cancer diagnostics and digital identity to industrial engineering and creative automation, companies across the Adria region are finding commercial applications for artificial intelligence. The question is whether the region can move beyond supplying technological expertise to owning the products, intellectual property and businesses that will shape the next digital economy.
The artificial intelligence industry has become extraordinarily good at selling the future. Every week brings another breakthrough, another billion-dollar investment or another prediction about professions destined to disappear. Yet amid the excitement, a more fundamental question is receiving surprisingly little attention: who is actually building businesses around AI, and who is simply spending money on it?
For the Adria region, the answer could have lasting economic consequences. Over the past two decades, Southeast Europe has established itself as a source of skilled engineers and sophisticated software development, supplying international companies with everything from enterprise applications to automotive technology. That success has created jobs, attracted investment and helped a generation of entrepreneurs build international careers. It has also exposed a familiar limitation: much of the value generated by regional expertise ultimately belongs to companies headquartered elsewhere.
Artificial intelligence offers an opportunity to change that relationship, although not necessarily in the way its most enthusiastic advocates imagine. The region is unlikely to challenge Silicon Valley in the race to develop the world’s largest AI models. Its more promising opportunity lies in applying advanced technology to specific industries, building proprietary products and solving problems for which international customers are willing to pay. Across the region, a growing number of companies are already pursuing precisely that strategy.
THE VALUE OF SOLVING REAL PROBLEMS
In Croatia, Microblink has spent years developing computer-vision technology that allows businesses to verify identity documents, automate customer onboarding and detect potential fraud. Its software is used in digital processes where accuracy, speed and security are essential, particularly in financial services and other regulated industries. In May 2025, the company reported that its platform had processed more than 12 billion documents from over 180 countries, a measure of the scale at which its technology operates.
Microblink’s significance extends beyond those numbers. At a time when generative AI dominates public attention, the company illustrates the commercial importance of less conspicuous applications of machine learning. Banks and other businesses do not need artificial intelligence to entertain customers with impressive demonstrations. They need systems that can identify people reliably, reduce fraud and make essential processes more efficient. The competitive advantage comes not simply from having an algorithm, but from developing the specialised technology, data expertise and customer relationships required to make it dependable.
The most valuable AI companies may not be those promising to change everything, but those solving problems businesses already pay to fix.
A similar principle applies to BetterMedicine, a medical technology company with Slovenian roots that develops AI-assisted diagnostic software. Its BMVision system is designed to help radiologists identify and assess kidney lesions in CT scans, addressing a problem facing healthcare systems across Europe: the increasing volume of medical imaging and the limited availability of specialists to interpret it.
The company reports that its technology has received CE marking under the European Union’s medical-device framework and has been deployed in hospitals in Germany and Estonia, with additional pilot projects in other markets. These achievements represent meaningful progress towards commercial adoption, although regulatory clearance and clinical validation should not be confused with evidence of widespread use or financial success. In medicine, the difference is particularly important. An impressive algorithm is only the beginning; the real test is whether it improves clinical practice safely, consistently and at a cost healthcare providers can justify.
Both companies demonstrate why the next generation of valuable AI businesses may look rather different from the technology giants dominating newspaper headlines. Their markets are specialised, their customers demanding and their products dependent on knowledge that cannot easily be replicated by simply connecting an application to a publicly available AI model.
THE OWNERSHIP QUESTION
The experience of HTEC, founded in Serbia in 2008 and now operating internationally, illustrates another dimension of the region’s technological ambitions. The company develops advanced engineering solutions for customers across industries including healthcare, automotive technology, semiconductors and industrial systems. Its capabilities extend into artificial intelligence, machine learning and complex software development, positioning it among the businesses helping international corporations incorporate advanced technology into commercial products.
HTEC’s expansion has also attracted significant institutional financing. In July 2025, the European Bank for Reconstruction and Development announced a €60 million loan to support the company’s acquisition-led growth. Earlier acquisitions, including Serbian engineering company SYRMIA, strengthened its expertise in areas such as embedded software and machine learning.
The transaction underlined the international potential of engineering businesses that originated in the region, even when their commercial activities now extend far beyond it.
Yet HTEC’s development raises a question relevant to the wider technology sector. Building sophisticated systems for international clients can create substantial economic value, but it is not necessarily the same as owning the products those systems support. The distinction matters because proprietary technology can generate recurring licensing revenue, stronger customer relationships and intellectual property whose value increases as a business expands. Engineering services can also be highly profitable and innovative, but their economics and ownership structures are different.
The challenge for the region is not to abandon engineering services, which remain an important source of employment, exports and technical expertise. It is to use the capabilities accumulated through those businesses to create more companies that control their own products and commercial futures.
In Bosnia and Herzegovina, Sarajevo-based Bloomteq represents a related approach. The company works on enterprise software, machine learning, custom large language models and AI-powered business applications, with activities spanning sectors including finance, healthcare, logistics and energy. Its broader ambitions include the Blum Institute for Artificial Intelligence, an initiative intended to connect technological development with research and specialist education.
Bloomteq has identified international collaborations involving organisations such as DFKI, Lenovo and Cloudflare, reflecting the importance of international networks for technology companies operating from smaller economies. What remains less clear from publicly available information is how much of its commercial activity derives from proprietary AI products rather than customised engineering services. That distinction is not a criticism of the company. It is precisely the kind of information needed to understand how the region’s emerging AI industry is developing.
AI ENTERS THE EVERYDAY ECONOMY
Not every commercially important application of artificial intelligence requires a scientific breakthrough. Some of the strongest opportunities arise when existing businesses use new technology to make familiar activities substantially more efficient.
Celtra, founded by Slovenian entrepreneur Miha Mikek, has built an international advertising technology business around creative automation, helping brands and agencies produce, adapt and distribute digital advertising across multiple formats and markets. Generative AI has become an extension of that business, offering new ways to accelerate creative production while maintaining consistency across campaigns.
In October 2025, Celtra published a case study describing how delivery platform Wolt had used its creative automation technology to produce more than 300,000 advertising assets in a year, with a reported increase in production efficiency of more than twenty times. Those results relate to Celtra’s broader automation capabilities rather than generative AI alone, but they illustrate something important about the commercial adoption of intelligent software. The strongest business case often comes from measurable improvements to existing operations, not from introducing entirely new activities.
The region has already proved it can supply world-class engineering talent. The next challenge is capturing more of the value that talent creates.
Celtra also benefits from an advantage that many AI startups lack: an established commercial platform, customer relationships and a clear understanding of the industry it serves. As AI tools become more widely available, those assets may prove as important as the technology itself. A company that understands how international advertisers operate can integrate new capabilities into services customers already use, while a newcomer may have to build both its technology and its market from scratch.
At a much earlier stage, North Macedonia’s Talbord is exploring how AI can simplify employee training and internal knowledge management. Its platform allows businesses to organise onboarding materials and use an AI assistant to answer employee questions based on company documentation. Support for Macedonian, Albanian and English addresses the practical requirements of a multilingual business environment that larger international software providers may not always prioritise.
Talbord says it has launched its product and is onboarding its first business customers. Its commercial future remains unproven, but the underlying opportunity is recognisable. Companies spend considerable time helping new employees understand procedures, locate information and become productive. A reliable system that reduces those costs could find a market well beyond North Macedonia, particularly among businesses facing similar organisational and language challenges.
BEYOND THE TECHNOLOGY ITSELF
These six companies represent different stages of development, different business models and very different levels of commercial maturity. Microblink and BetterMedicine are developing specialised technologies for demanding markets. HTEC and Bloomteq illustrate the opportunities available to advanced engineering businesses, while Celtra demonstrates how established software platforms can incorporate AI into services with identifiable commercial value. Talbord, meanwhile, represents the smaller ventures seeking to build businesses around practical applications of increasingly accessible technology.
Their differences reveal why counting AI startups or announcing national AI strategies tells us relatively little about an economy’s technological progress. The more useful questions concern ownership, customer demand, export potential and the ability to turn technical expertise into sustainable revenue. A company developing a proprietary medical diagnostic system faces entirely different challenges from a business integrating existing language models into employee software. Both may succeed, but they will require different financing, regulatory support and routes to international markets.
For the Adria region, the obstacles are not exclusively technological. Access to growth capital remains uneven, domestic markets are relatively small and ambitious founders often need to establish an international presence early in their development.
Skilled engineers are essential, but so are experienced commercial managers, international sales networks and customers prepared to adopt products developed by companies they may never have heard of.
Governments can help by improving research infrastructure, supporting collaboration between universities and industry, and making public procurement more accessible to innovative companies. Investors, meanwhile, will need to distinguish between businesses with genuine technological advantages and those whose products depend almost entirely on AI capabilities available to competitors.
Neither task is particularly glamorous, but both are more consequential than another conference declaration about the transformative potential of artificial intelligence.
The region’s strongest advantage may ultimately be its ability to combine technical expertise with an understanding of specific industries and their operational problems. Financial services, healthcare, industrial engineering, logistics and enterprise software offer opportunities in which specialised knowledge can matter more than the size of a company’s domestic market.
The companies that succeed will be those capable of converting that knowledge into products international customers trust and continue paying for.
There is no guarantee that the next generation of regional AI businesses will become global leaders. Some will remain successful engineering partners, others will be acquired, and many young ventures will struggle to establish commercially viable products. That is the reality of technology entrepreneurship, regardless of geography.
But the region has already demonstrated that its engineers can contribute to some of the world’s most sophisticated technological systems. The more consequential question is whether its entrepreneurs, investors and institutions can now translate that capability into businesses that retain more of the value they create.
The real measure of the Adria region’s AI ambitions will not be how enthusiastically it embraces the technology, but how successfully it turns that technology into businesses of its own.
