Banks don't struggle because they lack systems. They struggle because operational data arrives too late to matter.
A financial institution solved this with its Queue Management System, which tracked branch kiosks and service delivery but left operations teams reviewing month-old performance data in unreadable Excel files.
The API That Wasn't
The QMS offered FTP backup configuration with credentials, paths, and schedules. Promising on paper. In practice, the CSV exports had no headers, malformed data, and mixed incomplete datasets. Technically data existed. Practically, it was unusable.
Forcing this approach meant fragile parsing and low business trust. The project stopped.
The Pivot: Automate What Works
Instead of forcing machine integration, the team automated the human workflow the system supported well. The QMS had stable UI exports for user and service statistics as XLS files.
The solution: browserless Chromium automation running daily at 7:00 PM UTC+2. It logs in, navigates export screens, downloads two XLS files, and deposits them to SFTP /raw_data. No human action. Same structure every day.
Three-Layer Pipeline
Layer 1: Extraction (Daily, Reliable) Automated downloads deliver consistent XLS files to /raw_data.
Layer 2: Cleaning (Event-Driven) Webhook-triggered pipeline reads XLS, applies naming conventions, normalizes fields, calculates KPIs, converts to CSV, writes to /ready_data. Custom JavaScript handles business logic and data transformations.
Layer 3: Distribution (Scheduled) Cron job at 3:00 AM UTC pushes validated CSVs to external SFTP for Power BI consumption. Teams arrive to fresh dashboards reflecting current operations.
Tech Stack
n8n orchestration, custom JavaScript, browserless Chromium, SFTP zones, Ubuntu Server, Docker. Self-hosted. No core banking dependency. No vendor lock-in.
What Changed
Before: Monthly Excel exports, unreadable tables, DIY dashboards.
After: Daily automated data, current KPIs, decision-ready insights.
The lesson: The most reliable integration often respects how the system was designed to be used. When APIs fail, automate the stable UI workflow instead of forcing fragile parsing.
This mirrors broader APAC banking challenges where legacy systems lack APIs, forcing alternative integration patterns for operational analytics. Recent data shows monitoring automation can cut pipeline delays by 60% in liquidity reporting. The key is avoiding fragmented analytics that stay trapped in pilots without unified strategy.