Rocket Tech conducted this study in January–February 2026 through in-depth interviews with people in Kazakhstan. It examines customer experience in banking ecosystems and their related services, as well as demand for and willingness to adopt new services within these ecosystems, including AI agents.
The respondents were 10 people in Kazakhstan of different:
- gender (even M/F split);
- age (18–24 and 25+);
- experience with banking ecosystems (new users, long-time users and “refusers”).
- 10
- in-depth interviews
- 5
- user types
- 12
- ecosystem services
- 2026
- January–February
Key findings
- The main barrier. Cognitive friction: interfaces overloaded with services that cannot be turned off, interbank barriers and intrusive ads for unneeded services lower satisfaction across all segments.
- The leading ecosystem service. Kaspi’s public services and documents score highest among all ecosystem services and lead in satisfaction for 4 of 5 segments.
- Blind spots. Predictive AI analytics, children’s finance with educational features and notification settings, and integration with healthcare services are the least covered categories with huge growth potential.
12 parts of the study
Each part is a detailed review of one service: how different user types use it, a score from 0 to 10, respondent quotes, blockers and barriers.
Score map
How well banking ecosystems meet the needs of each user type in each service, scored from 0 to 10. On the right is the average per service, at the bottom the average per user type.
| Service | Normies | Pragmatists | Old-schoolers | Finance-savvy | Burned | Average |
|---|---|---|---|---|---|---|
| Payments and transfers | 9 | 6 | 5 | 7 | 4 | 6,2 |
| Food and grocery delivery | 8 | 7 | 4 | 6 | 5 | 6,0 |
| Fashion shopping | 8 | 7 | 4 | 6 | 5 | 6,0 |
| Transport | 6 | 8 | 4 | 7 | 5 | 6,0 |
| Entertainment and leisure | 5 | 8 | 3 | 5 | 4 | 5,0 |
| Financial planning and expense tracking | 6 | 6 | 3 | 8 | 3 | 5,2 |
| Travel | 5 | 8 | 2 | 7 | 0 | 4,4 |
| Public services and documents | 8 | 6 | 7 | 9 | 2 | 6,4 |
| Investments and brokerage | 8 | 5 | 9 | 5 | 3 | 6,0 |
| Children’s finance | 6 | 4 | 6 | 2 | 3 | 4,2 |
| Health and pharmacies | 5 | 8 | 2 | 3 | 2 | 4,0 |
| AI and chatbots | 3 | 5 | 1 | 6 | 1 | 3,2 |
| Average | 6,4 | 6,5 | 4,2 | 5,9 | 3,1 |
9–10 excellent7–8 good5–6 average3–4 weak0–2 critical
Five user types
Normies
Everyday infrastructure · average score 6.4
Use the ecosystem to support their everyday infrastructure. The bank anchors their familiar digital routine.
“A bank AI would be really convenient if it could answer any question about Kaspi… I’d even go to it instead of an operator”
AI readiness: moderate. Open only to the bank’s own AI. They want answers to questions from AI, not advice.
Pragmatists
Benefit optimizers · average score 6.5
Hunters for discounts and promotions. The bank is a tool for earning on their own spending.
“If the bank had its own AI… I’d use it for analytical queries, like which spending category went up”
AI readiness: skeptical. Only a transparent algorithm is acceptable. A “black box” instantly kills trust.
Old-schoolers
Low digital engagement · average score 4.2
A minimal set of tools. Any added complexity is seen as a barrier.
“I’m not tech-savvy, and I try to stay away from things I’d have to figure out”
AI readiness: low. Only an “invisible assistant” is acceptable. Pushing AI = the bank doesn’t want to solve problems.
Finance-savvy
Controllers and planners · average score 5.9
An analytical mindset. The bank is a data provider for their financial hub.
“When you track your expenses yourself, you understand them better… AI reduces your engagement”
AI readiness: high. They want AI as a “co-pilot”, not an autopilot. Ready to use it if control, security and transparency are ensured.
The Burned
Hurt and cautious · average score 3.1
Negative experience has left them with a “digital PTSD”. Reliability matters more than convenience.
“If the bank had its own ChatGPT, I wouldn’t really trust it… it’s a bank, they don’t specialise in AI stuff”
AI readiness: critical. Total distrust. They are sure that if something goes wrong, even a smart AI won’t help them.
The same study in the neighbouring country: Banking Ecosystems of Uzbekistan 2026.


