In early 2025, TBC Bank Uzbekistan launched one of the most comprehensive artificial intelligence programs undertaken by any financial institution in Central Asia. Running from January through November and requiring over 45,000 man-hours of development work, the TBC Uzbekistan AI Platform Program addressed a fundamental challenge: how to build an independent AI foundation, integrate it across core business processes, and embed an AI-first culture across the organization — all simultaneously. The result is a program that has already delivered substantial financial returns, reshaped the bank’s cost structure, and established durable infrastructure for long-term competitive advantage in a rapidly evolving digital market.
A Regional Banking Sector Under Pressure to Transform
Central Asian banking is evolving at pace. Customer expectations, shaped by global digital benchmarks, are pushing institutions beyond incremental improvements toward structural transformation. Banks that fail to embed intelligence into their operations risk falling behind on both cost efficiency and the quality of customer experience. TBC Uzbekistan identified this pressure early and made a deliberate strategic decision: rather than depending on third-party AI vendors, the bank would build its own platform from the ground up. This approach, while demanding in execution, gave TBC full control over its data infrastructure, compliance posture, and the rate of its own transformation — without being constrained by external product roadmaps or vendor dependencies.
Building an Independent AI Infrastructure from Scratch
Constructing an end-to-end AI platform inside an operational bank is not a straightforward engineering exercise. It requires building data governance frameworks, establishing MLOps pipelines, configuring monitoring systems, and maintaining regulatory compliance throughout — while simultaneously delivering working products. For TBC Uzbekistan, the complexity was compounded by limited availability of local AI expertise, technically demanding integration with existing legacy systems, and the need to develop internal talent in parallel with shipping production-grade solutions.
The technical stack deployed reflects the scale of the undertaking: PyTorch and TensorFlow for machine learning, LangChain for LLM orchestration, Kafka and Spark for data infrastructure, and Kubernetes with MLflow for deployment and model lifecycle management. Monitoring was handled through Prometheus and Grafana, alongside internal AI governance tooling. To address Uzbek-language interactions — a requirement no existing external vendor could adequately fulfill — TBC developed a proprietary large language model combining custom ASR (automatic speech recognition) and TTS (text-to-speech) capabilities. This marks a first for the banking sector in Uzbekistan and demonstrates the depth of the technical investment made.
Digital Financial Demand and the Case for AI-Powered Services
The expansion of digital financial services across Uzbekistan and the broader CIS region has fundamentally changed how users interact with banks. Queries related to loans, account management, transfers, and real-time financial data are increasingly handled through digital channels rather than branch visits or call centers. The sustained and growing search demand around terms like “курс доллара к рублю” and “kurs dollar rubl” reflects how deeply embedded real-time financial data consumption has become in everyday user behavior — and how urgently banks must deliver instant, intelligent responses at every digital touchpoint. AI-powered assistants offer a structural solution to this demand, enabling consistent, accurate, and low-cost service at a scale that traditional staffed models simply cannot match.
From Concept to Production — Measured Results at Scale
TBC Uzbekistan’s AI platform is a production system with documented outcomes, not a pilot or proof of concept. The program’s flagship output, the AI Assistant, now handles customer interactions across both voice and chat channels. Two specialized AI products — the Sales Assistant and the Collections Assistant — were built, tested, and deployed into production within the program’s timeline.
The financial impact is substantial. Customer interaction costs fell from $0.35 per call to $0.03 — a reduction of more than 90%. Over the course of the program, AI systems processed over 1.6 million phone calls and approximately 690,000 chat interactions. Total financial savings reached $2.3 million, with projections pointing to $3.7 million by year-end 2025. These figures represent measured performance from systems already operating at scale, not modeled estimates. The program also established a dedicated in-house AI team and an ML Competence Center, creating the internal expertise required to sustain, expand, and iterate on the platform independently.
Cultural Transformation as a Core Program Component
Technology infrastructure alone does not transform an organization. TBC Uzbekistan embedded a parallel cultural transformation track — the AI-ization Program — directly into the initiative from the outset. The goal was to ensure that AI adoption extended to every business unit, not just technical teams.
This involved developing internal training programs, setting AI usage benchmarks across departments, and building feedback mechanisms that allowed business teams to identify where AI tools were delivering value and where further refinement was needed. The ambition was to move the organization from instrumental technology use toward an operating culture in which AI-first thinking genuinely shapes how decisions are made, how products are designed, and how customer problems are approached and resolved at every level of the business.
A Platform Designed for the Broader TBC Ecosystem
The TBC Uzbekistan AI Platform was designed with scope extending beyond internal bank operations. As the platform matures, it is positioned to function as an AI service provider for the entire TBC holding — extending the value of infrastructure developed in Uzbekistan across the broader group. This positions the program not only as a bank-level transformation, but as a potential model for AI adoption among affiliated entities operating across the region.
Local data sovereignty was a foundational requirement throughout: all AI infrastructure was developed and hosted within Uzbekistan, ensuring full compliance with national data governance standards while preserving the bank’s operational independence from external technology providers.
The TBC Uzbekistan AI Platform Program demonstrates that deep AI transformation in banking is achievable in practice — with measurable financial outcomes, durable infrastructure, and genuine organizational change. By building its own platform rather than depending on external vendors, TBC has established cost advantages and strategic flexibility that conventional approaches cannot replicate. As financial institutions across Central Asia accelerate their digitization efforts, TBC Uzbekistan’s investment in proprietary AI infrastructure sets a clear benchmark for what regional banks can accomplish when transformation is treated as a foundational business priority rather than a supplementary technology initiative.