VBSP Smart Banking Redesign

◄ Back to projects

VBSP Smart Banking had a unique challenge: some banking tasks depended on people and processes outside the app, such as opening an account, which still meant a trip to a branch. This reinforced my view of HCI as understanding users’ goals and applying technology where it can meaningfully support them.

This redesign was the combination of user research, deep understanding of customer journeys, and the collaboration between Human & AI design.

New UI

Old UI

VBSP Smart banking redesign:

Financial journeys beyond the screen

Business Review and User Research

We sent surveys to VBSP to understand who their users are and where they

want the bank to be in five years. Those answers pointed us to the real pain

points and shaped the direction of the redesign. VBSP Smart Banking is meant

to be an instrument for financial inclusion, reaching more than 6.6 million poor

and near poor households across the country, so the main improvements to it

were suggested with the intent to serve their largest customer base, but also to close the gap between technological literacy for those who needed it most.
The deliverables for this project are confidential, but access can be requested to view, please contact me to clarify.

VBSP's three main goals

Goal 01

Make mobile banking easy for everyone

Goal 02

Reduce in-person transactions

Goal 03

Attract new customers in the private sector

How I built user understanding

Research was conducted under client confidentiality. Two current customers participated in moderated sessions. Identifying information, and recruitment materials are withheld. We also analyzed 259 public customer reviews, VBSP's policy documentation, stakeholder input via survey, and through informal user testing.

Client interviews

Two VBSP clients: an

elderly woman, and a

student who currently

borrows from the bank.

Desk research

VBSP policy, the 2030

strategy, the digital

transformation plan, and

annual reports.

Field input

Secondhand accounts

from Group leaders (Tổ

trưởng TK&VV), VBSP

staff, and families who use

the app.

Review mining

259 Google Play reviews,

read and tagged for

recurring pain points and

feature requests.

Partner survey

A requirements survey

filled out by NHCSXH on

their user base, needs,

and 5 year goals.

Translated into recurring usability themes ↓

Information Consistency and Hierarchy

Discoverability

Flexibility

Information load

Data Analytics and User Reviews

I scraped and classified 259 reviews from the Google Play Store to understand what users were actually asking for. I tagged each review by

sentiment, intent, and the Nielsen heuristic it pointed to, so a complaint could be traced back to a design cause.

259

reviews analyzed

Jun 2022 to Apr 2026

3.26

average rating

out of 5.0

15.4%

reviews asked for updated / new features

13.5%

reviews asked for online registration

Figure 1. Bar graph of positive and negative VBSP app sentiment (Green for positive (4-5 stars); Red for negative (1-3 stars)),

Figure 2. Pie chart of 1-5 star app store ratings (Top to bottom is marked 5 to 1 respectively)

Classifying by intent helped isolate and determine specific user needs. The most requested feature was online account registration as user's are still required to register in person at VBSP.

The highest tag for negative reviews (1-3 stars) was recognition over recall, stating that a lot of basic features were missing from the application. Other than large features such as online account registration and loan payment, many nesting component in main features (such as QR code scanning in the transaction feature) were missing a clear accessible pathway. We updated these features, changing the format for approximately 80% of the screens on the transcation feature, the home page, and the sim card purchasing feature.

Figure 3. Bar graph of Nielsen's heuristic evaluation violations (H1: System status · H2: Real-world match · H3: User control · H4: Consistency · H5: Error prevention · H6: Recognition over recall · H7: Flexibility · H8: Minimalist design · H9: Error recovery · H10: Help & documentation)

Figure 4. Pie chart of reviews based on user intent (From top to bottom respectively: Positive, Angry, New Feature wanted, Error, Help, Warning, Neutral)

Additional problems found from the application testing

Cognitive overload at the critical moment.

Confirmation and OTP were merged onto one screen -> User's were required to authenticate while confirming their transaction information

Inconsistent buttons, hidden fees.

Button position and wording changed from screen to screen, and service fees were not shown before the user confirmed.

Dead and hidden controls.

A language switch that appeared but did not work, no QR scan in the transfer flow, and buttons that did not work as intended.

The interviews, the review analysis, and the heuristic audit all pointed to the same three groups the bank leans on most.

Students
Many students from low income families borrow from VBSP to continue their studies. A student I interviewed studies in Hanoi, but her loan is managed back home in Hai Phong. She has to go through a lengthy process of transferring the money to her parents, have them take it out, and send a cash payment to the bank representative of their region.
What the user wants
From Hanoi, wants to pay her monthly student loan.
Opens the app.
Clicks on the module that leads them to the loan payment feature.
Pays their loan
What the user is experiencing
From Hanoi, wants to pay her monthly student loan.
Opens the app.
Sees their loan balancing but can't access a loan payment feature
Has to contact someone from their hometown, transfer the money
to them, and ask them to take the money to a bank representative.
In her words: “I would rather pay the bank myself by internet transfer than route the money through my parents and the group leader every month.”
Key finding: For students living away from home, repayment is not an app task, it is an offline relay through parents and a group leader → a direct online payment would collapse a multi person, 10 km errand into a single transfer.
Farmer / Agricultural labor force
About 80% of VBSP's customers rely on policy programs, and the largest group among them is farmers and agricultural
workers, many of them women. They mostly open the app to track a loan or reach the bank, and what they asked for was
simple: a friendlier, more accessible interface, more ways to pay a loan back, and a way to follow their own spending, instead of
always going through a regional bank representative.
What the user wants
Wants to open an account and start borrowing.
Opens the app.
Accesses loaning information from bank.
Contacts the bank and starts their loan.
What the user is experiencing
Wants to open an account and start borrowing.
Opens the app.
Can't sign up for a bank account online.
Has to travel to a bank branch in order to open an account
Key finding: A online account registration feature would cut the demand for in person meetings and speed up the time needed to use bank services → it answers the user's need and the bank's at the same time.
Elderly customers
VBSP serves many older customers, and low vision is often the real barrier, not a lack of experience. The woman I interviewed is
56, lives in Hanoi, and already uses Agribank and Vietcombank without help, so her struggle was never about
finding her way around a banking app. It was about this one.
What the user wants
Wants to send a transfer on her own, the way she does
on her other banking apps.
Opens the app and finds the transfer button right away.
Enters the details at a size she can read, zooming in if needed.
Confirms with her PIN, and she is done.
What the user is experiencing
Spends over a minute just finding the transfer button, hidden in the header image.
Types in the recipient account information (3 minutes)
Tries to zoom in on the confirmation page, but can’t.
Finishes transaction but takes a long time to get there.
Observed transfer run: 1 min 3 sec to find the transfer button, 3 min 26 sec to enter the details, 37 sec paused on the confirmation page, 32 sec on the PIN. She then asked how to zoom in, the screen never responded to her zoom motion.
Key finding: She already banks on two other apps, so this was not about understanding banking flows. The real problems were button visibility, low contrast, lack of tooltíp and no response to phone's native accessibility features.

Wireframing faster, and more sustainably

As the number of flows grew, manual wireframing became a bottleneck. I developed an AI-assisted pipeline that generated editable wireframes from structured requirements, while keeping review and revision under designer control. What began as a workflow experiment during VBSP later became the foundation for my design department's AI Team's (which I am a member of) Open Design project.

That human in the loop setup later grew into a standalone pipeline of its own. The full build, the thinking behind it, and the points where I keep

a hand on the design all live in my Open Design UX Research pipeline.

Impact and Lessons

Once the design was ready, we handed it to the development and testing teams.

✓

20% faster money transfers on average compared with the existing flow, based on a TestFlight test run by the testing team. This excludes system processing time; the rest of the test details are confidential.

✓

70 to 80% less in person traffic (the client's projection for the new online registration flow).

✓

Built to WCAG guidance, with extra care for contrast, target size and readability for older and low vision customers.

The real problem usually lived off the screen. Designing for a policy driven market, so unlike consumer banking, was the hardest part and taught me the most. It really challenged me to think outside of the box, spend a lot of time on user interviews and feedback and worked closely with the bank to understand their specific work flow.

It was also one of my first times building AI into my process. Keeping a human in the loop and counting what the work consumed made it faster and more repeatable. That lesson started my pipeline, AI in Design.

The New Design

With the findings from research and the structure from the wireframes, we built the new screens, keeping clear and reachable interaction design at the center of every decision. We leaned on the banks three main goals: accessibility, reduction in in-person transactions, and increase in private sector customers. By updating basic features and adding ones of interest for target customers (more personalized donation feature and faster bank transfer flows), we were able to recommend ideas and changes that the bank was looking for.

This design is a place holder for our team's UI work due to the confidentiality of the project. These designs were still designed by me and hold the logic for our work.