By Meelis Kitsing
Estonia’s digital success shows AI transformation depends not just on technology, but on connected systems, better organization, and willingness to experiment beyond old structures.
The future trajectories of transformation driven by artificial intelligence (AI) must be addressed in the plural. Profound uncertainty makes it difficult to determine the extent to which AI will prove complementary or substitutive. The potential substitution effect itself increases uncertainty. It is easier to identify where AI may replace existing processes and human endeavors. However, its transformative potential is far more difficult to imagine in areas where AI could create entirely new realms of activity.
This is part of a broader story of technological progress in which past experiences offer valuable insight. Estonia’s digital transformation over the last 30 years has often been portrayed primarily as a story of efficiency. A small country with limited resources built one of the world’s most advanced digital societies through bold experimentation. Yet the deeper lesson of Estonia’s experience is not merely technological but organizational. Digitalization made life easier for people. Estonian organizations did not simply digitize existing processes; they created new systems that were simpler and more transparent. Both public and private organizations in Estonia understood early on that value in the digital age emerges not from centralized control, but from networks, interoperability, and ecosystems.
These insights are empathized by prominent British innovation expert and serial entrepreneur David Cleevely in his recent book Serendipity which he discussed in a seminar at the Estonian Business School in May this year. When Estonia developed the X-Road architecture, it rejected the idea that all data should be centralized into one massive state database. Instead, X-Road created a decentralized interoperability layer that allowed different institutions to exchange data securely while maintaining autonomy. This architecture was not merely technical elegance. It was a governance philosophy. Agencies retained ownership over their systems, but interoperability enabled ecosystem-level value creation. In some ways it ways blockchain before this term became well known.
Cleevely also highlights the examples of Kazaa and Skype, which were developed by the same team. Skype emerged from a decentralized peer-to-peer architecture rather than from centralized telecommunications infrastructure. Its success demonstrated how networked ecosystems could disrupt large incumbents by reducing friction and empowering users directly.
I would emphasize that networks are not just about technology but, above all, about people. What I call the “Skype effect” indirectly inspired a new generation of entrepreneurs and directly created a cohort of future unicorn founders through experience, networks, and funding. Estonian unicorns Bolt and Wise are cases in point, as some of their co-founders previously worked at Skype.
Even before Skype and X-Road, Estonia’s internet banking revolution had begun thirty years ago. Estonian banks embraced digital identity, interoperability, and user-centric services earlier than most European competitors. What mattered was not simply digitizing paperwork, but redesigning the customer experience around trust and seamless interaction. Estonian banks entirely skipped some legacy services, such as paper checks, which are still widely used in many countries.
Most importantly, internet banking platforms became gateways for public services. In 2000, the Estonian Tax Board moved tax declarations online using identification provided by private banks. Estonia’s public ID card system, usable both online and offline, was launched a few years later.
Today, many organizations in adopting AI remain stuck in a familiar pattern: isolated pilots, fragmented data projects, and disconnected tools that fail to scale. Let’s take a look banking in order to build on the Estonian digitalization experience. Banks deploy chatbots, automate compliance tasks, or experiment with fraud detection, but struggle to transform their overall operating model. The challenge is not the absence of algorithms. It is the persistence of silos.
The current wave of AI adoption in banking resembles in some ways the early internet era. Most banks have digitized interfaces but still operate internally as fragmented organizations. Customer data sits in separate silos across retail banking, corporate banking, wealth management, payments, compliance, and insurance. AI systems are frequently layered on top of these fragmented structures rather than transforming them because they cannot use contextual data. One department often cannot effectively use information generated by another. This limits AI’s strategic potential because intelligence depends on connectedness.
As a result, banks often automate inefficiency, as was done before by digitalization of processes associated with checks, instead of redesigning value creation. The more profound transformation will occur when banks shift from siloed institutions toward AI-enabled financial ecosystems.
I often conduct scenario planning workshops about AI-driven transformation in organizations. AI regulation and AI advancement emerge as prominent key drivers in these discussions among others. Since AI is a moving target regulation is not just about written text, but ability to implement the rules in practice. Regulatory cycles are slow while technology are fast. Having said that AI advancement is not given. Innovation is about using inventions in practice which implies organizational as well as broader social acceptance of AI.
One potential future trajectory for banking resembles Estonia’s X-Road philosophy. The future bank may no longer function as a closed institution controlling all services internally. Instead, it may become an orchestrator within a broader ecosystem involving fintech firms, digital identity providers, cloud platforms, AI developers, payment networks, regulators, cybersecurity providers, and even non-financial platforms. Interoperability becomes more important than ownership. In this scenario “Banking without banks”, the real competitive advantage in AI-driven banking will therefore come from superior ecosystems. Financial services could evolve into decentralized service ecosystems where customers dynamically combine services from multiple providers.
Yet the greatest barriers for such AI-driven transformation scenarios are not technological. They are organizational, political and regulatory. Banks remain heavily hierarchical institutions shaped by regulatory caution and legacy infrastructure. Compliance concerns discourage experimentation. Incentive structures reward risk minimization rather than ecosystem innovation. Many institutions still approach AI as an IT project rather than a strategic redesign of value creation.
“…the innovator has for enemies all those who have done well under the old conditions, and only lukewarm defenders among those who may do well under the new,” wrote Niccolò Machiavelli in The Prince. This captures a classic dilemma of the creative destruction associated with innovation: it is easier to see what may be lost than to imagine what could be gained. Estonia’s digital transformation suggests that we should not allow our collective imagination to fail us and that the best path forward lies in bold experimentation.
Meelis Kitsing
