RAMS — Registration Asset & Materials Management System
A Model-Driven Power App that became the single source of truth for registration stock, supply, procurement and vendor management across the Cox’s Bazar office and eight registration sites.
Screenshots of live humanitarian systems are not published. Where visuals appear, they are illustrative rebuilds created with synthetic data that reproduce the structure and design of the original work without exposing any operational or personal information — on data-protection grounds.
Context
Registration operations across the Cox’s Bazar office and eight registration sites depend on a constant supply of consumables and equipment — ID card stock, printers, biometric hardware, forms. Tracking sat in disconnected spreadsheets, so stock levels were only known after someone physically went and checked.
The consequence was not an inventory problem, it was a service problem: when a site ran out of card stock, the registration desk stopped. People who had travelled to be registered were turned away. Procurement made it worse — without a reliable stock position, reorder requests were raised late and against guesswork.
Approach
I designed and built RAMS in 2025 as a Model-Driven Power App on Dataverse — a single source of truth spanning stock, supply, procurement and vendor management rather than inventory alone.
- Why Model-Driven, not Canvas. Assets, sites, movements, purchase requests and vendors are a genuinely relational problem with approval chains attached. Dataverse gives that a proper data model, role-based security and a native audit trail as standard — all things a Canvas app would have needed hand-building, badly.
- One model, four functions. Stock (what is held, where), supply (what moves between office and sites), procurement (what is on order and at which stage), and vendor management (who supplies what, on what terms) share a single schema instead of four spreadsheets that disagree.
- Automation. Power Automate flows for low-stock alerts and approval routing — the system raises the flag before a site notices, and a reorder follows a defined path rather than an email chain.
- Visibility. A Power BI layer over Dataverse for management oversight across all nine locations at once.
How it looks
Asset value by category
Rebuilt with synthetic values, real category structure
Asset value by location
Synthetic values — the office and eight registration sites
Procurement pipeline
Synthetic volumes — the stage structure is the real one
Outcome
US$4M equivalent under management
RAMS is in active use, managing registration assets worth US$4M equivalent across the Cox’s Bazar office and eight registration sites. Stockouts were eliminated.
The wider change is that four separate activities — stock, supply, procurement and vendor management — now run off one record rather than four spreadsheets. Asset accountability moved from periodic manual counts to continuous visibility, and reordering moved from reactive to scheduled, with an approval trail attached to every request.
What I’d do differently
The reorder thresholds are static values set per category. They should be dynamic — derived from each site’s actual consumption rate, which varies a great deal with the exercise running at the time. That is a forecasting problem, and it is the natural next version.
Vendor performance is recorded but not yet scored. Lead time and fulfilment accuracy are both in the data already; turning them into a simple supplier rating would make the procurement decision evidence-based rather than habitual.