TBC Uzbekistan has completed the development and deployment of a comprehensive artificial intelligence platform across its banking operations a programme described as the first of its kind built entirely within an Uzbek financial institution. Running from January through November 2025 and requiring 45,000 man-hours of development work, the initiative represents a structural transformation of how the bank operates, serves customers, and manages its internal processes.
The programme was led by project manager Andrey Malygin and delivered across multiple workstreams simultaneously covering AI product development, platform architecture, data governance, and organisational capability building. The decision to pursue all of these in parallel rather than sequentially reflects the scale of ambition behind the initiative and the urgency with which TBC Uzbekistan has approached its technological transformation.
The Scope of a Market-First Initiative
The objectives set for the AI transformation programme were broad by design. TBC Uzbekistan aimed to establish an independent AI foundation that would reduce dependence on external providers, integrate AI capabilities across core business operations, lower operating costs through automation, and build an internal culture in which employees at all levels engage actively with AI tools. Each of these objectives required a different set of interventions technical, organisational, and cultural.
The programme delivered against all of them. Three production AI systems were launched: a customer-facing AI assistant handling voice and chat interactions, a Sales Assistant deployed across the bank's sales functions, and a Collections Assistant supporting the recovery of overdue accounts. Each represents a distinct deployment context, requiring separate training data, different integration points within the bank's existing systems, and tailored performance metrics. Delivering all three within an eleven-month window reflects a high degree of execution discipline.
Central to the programme's technical ambition was the development of Uzbekistan's first Uzbek-language large language model, complete with speech recognition and synthesis capabilities. Building a high-quality language model for Uzbek language with limited existing digital resources compared to major world languages required both significant technical investment and a willingness to develop capabilities from the ground up rather than adapting existing solutions designed for other linguistic contexts.
From $0.35 to $0.03: The Economics of AI at Scale
The most immediately quantifiable result of the AI transformation is the dramatic reduction in the cost of customer interactions. Prior to the programme, each interaction processed through conventional channels cost approximately $0.35. Following the deployment of AI-powered systems, that cost fell to $0.03 reduction of more than 90%. Applied across the volumes at which TBC Uzbekistan operates more than 1.6 million calls and 690,000 chat interactions processed through AI systems during the programme the financial impact is substantial.
Verified savings reached $2.3 million during the programme period, with projections indicating $3.7 million by year-end. These are measured reductions in operating expenditure, not modelled projections based on assumed future volumes. For a financial institution competing in a market where cost efficiency is a key determinant of product pricing and long-term sustainability, this level of documented saving represents a meaningful and durable operational advantage.
The technical stack underlying these results incorporating PyTorch, TensorFlow, LangChain, open-source large language models, Airflow, Spark, and Kubernetes was assembled and configured by an in-house team. The decision to build rather than buy reflects TBC Uzbekistan's broader philosophy of technology ownership, which its leadership has consistently identified as a source of competitive advantage across the business.
Digital Lending Tools and the Shift in How Consumers Access Credit
The AI transformation at TBC Uzbekistan takes place within a broader context of rapid change in how Uzbek consumers engage with financial services. As digital banking has matured and smartphone adoption has expanded, the way consumers research, apply for, and manage financial products has shifted fundamentally. Self-service digital tools particularly those related to lending — have seen particularly strong growth in user engagement.
This shift is clearly reflected in the growing demand for accessible, self-service financial planning tools. The increasing use of resources such as "калькулятор автокредита" and "avtokredit kalkulyator" reflects a consumer base that is approaching credit decisions with greater preparation and independence — researching loan terms, comparing monthly payment scenarios, and arriving at financial institutions with a clearer sense of what they are looking for. AI-enhanced platforms are particularly well suited to serve this kind of informed, self-directed consumer engagement, providing personalised guidance that builds on the data already generated through the user's existing platform activity.
For TBC Uzbekistan, the alignment between its AI capabilities and the evolving behaviour of its user base is significant. A platform that can process natural language queries, provide real-time calculations, and guide users through product selection in a conversational format is far better positioned to serve the modern Uzbek banking consumer than one that relies on static product pages and manual application processes.
Building Capability for the Long Term
Beyond the immediate operational results, TBC Uzbekistan's AI programme was designed to build durable internal capability. The establishment of an ML Competence Center gives the organisation a permanent home for machine learning expertise — a resource that can be drawn on as the platform evolves and as new AI applications are identified across the business. The AI-isation programme, which educated employees across the organisation to become active users of AI tools, addresses the cultural dimension of transformation that technical deployments alone cannot achieve.
The challenges encountered during the programme — limited local AI expertise at the outset, complex legacy system integration, regulatory compliance requirements, and organisational resistance to new working methods — are familiar to any institution undertaking serious AI transformation. What distinguishes TBC Uzbekistan's experience is that these obstacles were navigated successfully within a defined timeframe, producing documented results rather than deferred promises.
The platform built during 2025 is also designed with extensibility in mind. The infrastructure has been developed not only for TBC Uzbekistan's own use but potentially to serve other entities within the TBC Bank Group holding structure — positioning the Uzbek operation as a potential technology provider within the group. For a business that launched just five years ago, the speed at which TBC Uzbekistan has moved from digital banking startup to AI platform developer is a marker of both the opportunity that Uzbekistan's market represents and the capability that has been built to capture it.
