Kazakhstan's fintech has accustomed users to solving any task at the push of a button. However, the market is now on the verge of transitioning from a "set of buttons" to intelligent dialogue. While global giants are implementing advanced large language models (LLMs), local banking ecosystems still rely on traditional scripted chatbots. This creates a noticeable gap: users are already used to solving work and personal tasks via ChatGPT or Gemini, but inside banking apps, they still encounter "digital answering machines."
What is this research about?
In this research, we conducted in-depth interviews with Kazakhstanis to study their experience interacting with banking ecosystems.
In this article, we analyze services with artificial intelligence:
- Why do advanced users leave banking apps for external neural networks to analyze their own finances?
- How do different segments perceive the replacement of a live operator with an algorithm?
- What barriers prevent AI from becoming a full-fledged financial assistant inside a smartphone?
Below is a detailed breakdown of five archetypes defining the landscape of intelligent services in Kazakhstan.
😐 Normies
For them, a banking app is a utilitarian tool where everything must be perfectly clear. AI is valuable to them only when it eliminates the need to search for information or wait a long time on the support line.
Artificial intelligence and chatbots (Rating: 3/10)
The needs of this segment are minimally met. Existing bots (like "Aidos" in Halyk) seem too primitive to them—they rather direct to an operator than solve the problem.
Quote: "Banking AI is very convenient, if it could answer any question on Kaspi... I would even go to it instead of an operator."
Behavior: They turn to bots only in extreme cases when they do not want to wait for a human response. They expect simple document prompts or automation of standard payments ("pay utilities like last time") from AI.
Blockers and barriers
- "Dead-end scenario": The main unique blocker is the limitation of answers. If the bot does not understand the request and starts going in circles ("I didn't understand you, choose a category"), the normie instantly loses trust and demands a live operator.
- Difficulty of access: If the chatbot is hidden deep in the support menu, the normie simply does not reach it, preferring to act the old-fashioned way.
- Disconnect from personal data: Switching to external neural networks (ChatGPT/Deepseek) to search for information instead of banking AI due to the lack of connection between "smart" search and their account status.
🤑 Pragmatists
Rational users who already use external AI (ChatGPT) to compare benefits. They expect the bank to offer them a built-in "intelligent calculator" that will manage their money itself.
Artificial intelligence and chatbots (Rating: 5/10)
The average rating is due to the huge gap between their expectations and reality. They see the potential in AI for automatic budget optimization, but in apps, they find only static analytics.
Quote: "If the bank had its own AI... it would be useful for various analytical queries—for example, which expense category has grown. It could help with balance notifications and budget planning."
Behavior: They use neural networks outside banking apps to calculate deposit benefits or compare cashbacks. Inside the bank, they are "starving" for real recommendations that would save them money.
Blockers and barriers
- "Static data": The main blocker is the lack of proactivity. The AI in the bank sees the spending but gives no advice: "This month you overpaid 10,000 KZT for taxis, use bonuses." Without such insights, AI is useless to a pragmatist.
- Lack of "autopilot": Pragmatists are ready to entrust AI with automatically transferring money to a deposit with the best rate, but the lack of such a feature makes banking bots look outdated in their eyes.
- External benefit arbitrage: Conscious search for promotions and price comparison through external AI tools, as banking AI is not perceived as an objective assistant in saving money.
🥸 Olds
A conservative segment for whom any AI is a barrier between them and the bank. They perceive chatbots as a way for the bank to save money on live people at the expense of customer convenience.
Artificial intelligence and chatbots (Rating: 1/10)
The lowest rating. For them, AI is synonymous with an "incomprehensible thing" that only complicates life and deprives them of human contact.
Quote: "AI support, unfortunately, does not yet reach the level of the simplest operator. I prefer to speak with a human."
Behavior: They either ignore chatbots or immediately type "Operator" in the chat to bypass the algorithm. They experience deep distrust in the reliability of AI answers in financial matters.
Blockers and barriers
- "Digital wall": A unique blocker is the feeling that the bank is "hiding" behind the bot to avoid solving the problem. To the olds, this looks like disrespect for the client.
- Incomprehensibility of the interface: Voice assistants or chat boxes often require "correct" query phrasing, which is an insurmountable barrier for the olds.
- Distrust in data security: Deep fear for privacy—it is unclear to the olds where their queries "go" and who gets access to financial information after communicating with AI.
🧐 Financially Literate
Experts who have already integrated neural networks into their work and investments. They expect a "financial copilot" from the bank that will work with their portfolio at the level of a professional analyst.
Artificial intelligence and chatbots (Rating: 6/10)
They see the beginnings of the necessary tools in banks (brokers in BCC, Freedom), but the lack of intelligence in analytics makes these services merely "information boards."
Quote: "I am currently studying the stock market using ChatGPT... analyzing charts with it. It would be useful to have an AI assistant on the phone directly for finances—for reconciling statements and planning."
Behavior: They upload screenshots of charts and statements to external LLMs (Claude, ChatGPT, Perplexity) to get quality analysis. They use AI to generate diversification ideas, as banking bots are not yet capable of this.
Blockers and barriers
- "Information vacuum" in analytics: The main blocker is the lack of deep LLM tools inside banking reports. It is not enough for them to know "how much was spent"; they need to know "why" and "what will happen next" in the format of a live dialogue with data.
- Lack of access to market insights: A blocker is that the banking bot cannot analyze external market news and correlate it with the user's personal portfolio.
- Open banking barrier: The desire to get quality AI analytics while maintaining an acute fear of granting the algorithm access to data from other banking accounts.
😡 The Burned
People with negative experiences who see AI not as an assistant, but as a new point of failure. For them, an algorithm is an entity with which it is impossible to negotiate or hold accountable in case of an error.
Artificial intelligence and chatbots (Rating: 1/10)
Critical distrust. Past bank errors (incorrect debts, blockages) make them perceive AI as an additional risk factor.
Quote: "I encountered chatbots in banks—it didn't resolve that situation with the debt for me, I had to call a human. I wouldn't want anyone to have access to info about my money besides me."
Behavior: They consciously restrict apps' access to their data. Any offer to "try an AI assistant" is perceived as an attempt by the bank to reach even deeper into their pockets.
Blockers and barriers
- Fear of algorithm "hallucinations": A unique blocker is the fear that AI will "invent" a debt for them or misinterpret a transaction, leading to a blockage that will be impossible to appeal.
- Distrust in the bank's competencies: They believe the bank should deal with money, not neural networks. "AI from a bank" sounds to them like a dubious and dangerous experiment.
- Total refusal: A principled refusal of any AI functionality inside the bank after past systemic errors of the organization or negative experiences with support.
Final summary
At the moment, AI in Kazakhstan's banking ecosystems is a "hungry" market. The most valuable segments (Pragmatists and Financially Literate) are already ripe for intelligent assistants and are forced to use external tools while banks offer them primitive bots.
The main insight: trust in AI will not arise from chats; it will arise from utility. The ecosystem that wins will be the one that offers not a "conversationalist," but a closed, secure analytical algorithm. It must be able to work in a "my data only" mode, offering real scenarios for saving and capital growth without an intrusive interface and the risk of "hallucinations."
