External AI tools like ChatGPT and Gemini have effectively become the primary search engines for solving complex, multi-component tasks for Uzbekistanis. However, there is nothing comparable available within banking ecosystems.
There is an acute, unmet demand in the market for embedded financial AI that could compare products, suggest benefits, and analyze spending. At the same time, the main brake on the technology's development is a colossal barrier of distrust: users are categorically afraid to give algorithms access to their personal financial data.
What is this research about?
In this study, conducted by Rocket Tech in January–February 2026, we held in-depth interviews with Uzbekistanis to explore their experience interacting with banking and payment ecosystems.
In this article, we analyze information search and attitudes toward AI:
- How do different user segments apply neural networks in their lives?
- How do they envision the ideal artificial intelligence in a bank?
- What are they missing, and what stops them from granting AI access to their finances?
Below is a detailed breakdown of four archetypes that define the landscape of Uzbekistan's fintech.
🌐 Ecosystem Nomads
Active users, technically literate, 25+, working in IT or related fields. They perfectly understand how technologies work and clearly formulate their requests.
Search and information processing (Score: 4/10)
They actively use external AI for work tasks, business ideas, and price comparisons, but are disappointed by the complete lack of such tools in local banks.
Quote: "It would be cool if there was a neural network that knows the products of all banks and microfinance organizations—comparing and selecting through it would be easier than going through them manually" (Azamat).
Behavior: Forced to manually collect data from bank websites and upload it to ChatGPT for analysis and financial recommendations.
Blockers and barriers
- Lack of localization: Global AIs do not know the specifics of Uzbekistan's banking products and local legislation.
- Inability to cross-compare banks: The lack of a tool capable of parsing and comparing competitors' terms in real-time within a single window.
- Disconnect between search and action: After finding a profitable product or deposit via external AI, the user has to perform many manual actions to apply for it.
🛡️ Cautious Functionalists
Mostly women 18–25+, use apps for specific tasks, value stability and trust more than innovation. Skeptical of AI in finance.
Search and information processing (Score: 4/10)
They use ChatGPT for everyday queries (deciphering recipes), but treat financial advice from an algorithm with strong reservation.
Quote: "Regarding AI access to my data, I would be wary—it's unclear what could happen in terms of spreading info about my finances" (Visola).
Behavior: Refuse to integrate their data with smart assistants, demanding clear legal guarantees before any interaction.
Blockers and barriers
- Fear of data leaks: Panic-driven distrust of how and where financial histories will be stored if a neural network gains access to them.
- Legal opacity: Lack of clear agreements on who bears responsibility for AI errors or data compromise.
- Deficit of trust in the algorithm: The belief that AI is incapable of taking subtle life circumstances into account during financial planning.
🌱 Youth
Young users (18–25+), just entering the ecosystem. Open to new things, actively integrating technologies into their daily lives.
Search and information processing (Score: 5/10)
Use AI as actively as possible (in studies, coding, everyday life) and have the lowest barrier of trust toward implementing AI in a bank.
Quote: "I don't see any negative sides to a banking AI having access to my banking data—after all, the bank already has access to it anyway" (Sardor).
Behavior: View AI as a universal advisor and search engine, replacing long Google surfing.
Blockers and barriers
- Lack of a built-in AI search engine: The inability to find a needed service in the bank using simple human language (when you forget the "correct name").
- Fragmented user journey: The need to exit the bank to Gemini/ChatGPT to clarify an obscure financial term.
- Weak interfaces of current apps: Current search systems in banks do not understand context and only work with exact keywords.
🙅♂️ Rejectors
Users who have significantly reduced their use of superapps due to interface overload and negative experiences. Prefer "narrow" tools.
Search and information processing (Score: 2/10)
Hardly use AI tools, and feel maximum caution and skepticism toward banking neural networks.
Quote: "Regarding AIs in financial systems—I wouldn't give access, I don't trust these models that much yet" (Abduvakhid).
Behavior: Tried chatting with ChatGPT but were disappointed in its usefulness, preferring to solve all financial issues "in their head."
Blockers and barriers
- The "yes-man" effect: Perceiving AI as a useless toy that simply agrees with the user and provides no real expertise.
- Fear of losing control: The belief that full algorithm access to data will lead to automatic erroneous write-offs.
- Terminological rejection: The very phrase "AI in a bank" is perceived by them as a sign of instability and danger.
Final summary
Search and information processing using AI is an area of colossal pent-up demand from a technically advanced audience. However, the large-scale implementation of AI assistants in Uzbekistan's banking apps is impossible without proactive work on the objections of conservative users, for whom data security and privacy are more important than any convenience.
