An internal tool for bank staff to collect customer documents over the counter. Covers photo capture, video recording, document scanning, and signature collection.
The developer built the app before I was brought in. My job was to design around what already existed, not start from scratch. The brief was to make it work well for bank staff using it over the counter, all within a week.
Tellers hand the phone to customers to capture documents during counter visits. The app needed to support four capture modes: photo, video recording, document scanning, and signature. Everything collected goes to the bank's library storage.
The main screen was the hardest to get right. With just a list of task buttons and nothing else, it kept looking flat and dated. Everything had the same visual weight and there was nothing to anchor the layout. The hierarchy felt off.
Four parts of a teller's day: starting a session, signing in, capturing documents, and checking what was collected. Every screen works in light and dark mode. Click any image to zoom in.
Adding a hero card at the top of the main screen fixed the balance problem. It shows the bank counter, teller name, and date. This gives the screen context before the task buttons and creates a natural focal point that made the layout feel intentional. Every screen works in light and dark mode.
The auth flow covers login, sign up, a security phrase check, and a full password recovery path. Password error handling mattered a lot here since this is a staff-facing app. Tellers need clear feedback when something goes wrong mid-handoff.
Each capture mode follows the same structure: capture, preview, confirm or retry, then upload. Keeping the pattern consistent meant staff could switch between modes without having to relearn the flow each time.
After uploading, staff can review what's been captured in the storage view and preview files before the session ends, so anything that needs a retake gets caught. Settings covers switching the app's language, logging out between shifts, and requesting data deletion. Language support mattered since tellers serve customers in Lao, Mandarin, and Thai as well as English.