Elevator servicing · Ghana
From a fully manual operation to a company you can query.
ARG
The problem
ARG services elevators across ‹N› buildings, and until this system the whole operation ran manually. Equipment histories lived in engineers' notebooks and one shared spreadsheet. Nobody could answer, quickly and reliably, what was installed where, what had been replaced, or what a job had actually cost in parts. That made maintenance planning and client billing guesswork.
What we built
- Equipment registry with full lifecycle history
- Stock management with last-price-in revaluation via database RPCs
- 16-report analytics suite
- AI-generated preventive maintenance scheduling from equipment history
- QR-coded asset tags for on-site scanning
- Per-user authentication with HMAC-signed sessions
- Repair and maintenance service log
Architecture notes
Equipment is the architectural DNA. Every stock movement, service visit and report resolves back to a piece of equipment, which is what makes cost-per-asset answerable at all. Multi-tenancy was built in from day one with tenant scoping on every table, so ARG's second entity is a configuration change rather than a rebuild. The system shares a database with a legacy application still in daily use, so every schema change is strictly additive and migrations ship as committed SQL files. The legacy app has never broken during a deployment.
Outcome
Phase 1 in production. Every job, part and visit now lands in a database the company can query, so decisions that used to run on memory run on data the system gathers as a side effect of daily work. Stock valuation that used to take a month-end reconciliation is now current. Phase 2, asset tagging and service logging, is scoped and underway.