Most businesses outgrow their spreadsheets and shared inboxes long before they admit it. We build the platform that replaces them — designed around your actual process, owned by you, maintained by the people who wrote it. We do exactly the same for our own product.
Every packaged product makes an assumption about how you work. When that assumption is close enough, buy it — we will tell you so, and we have. The projects worth building are the ones where the mismatch is quietly costing you hours: the reconciliation someone redoes every Friday, the WhatsApp group that has become a job queue, the three systems that disagree about the same invoice. Custom software is worth it when the process is the business, not when it is merely inconvenient.
We build those systems, and we run one of our own. Cargonio is our product — a multi-tenant freight-forwarding ERP covering air and sea import-export operations, invoicing, PRA and FBR e-IMS tax filing, double-entry accounting, HR, payroll and over a hundred reports. It is not a demo built for a portfolio. It has customers, a release cycle, migrations that have to run on live data and support tickets we answer ourselves. Everything that taught us goes straight into client work.
Job files, approvals, scheduling, stock, commissions — the work your team currently tracks in Excel and chat threads, moved into one system with roles, permissions and an audit trail of who changed what. Most of our builds start here, because this is where the pain is measurable.
Cargonio gave us the awkward parts in production: data separation between tenants, per-tenant settings and branding, subscription billing, and shipping a change to everyone at once without breaking the customer who configured things differently. If you are selling software into your own industry, that is the same problem set.
PRA and FBR e-IMS fiscal invoicing in Pakistan, MyFatoorah in Kuwait, Stripe for card payments elsewhere. The interesting work is not the happy path — it is retries, reconciliation, idempotency, and deciding what the user sees when a government endpoint is down for the afternoon.
Double-entry ledgers that actually balance, PDF invoices and statements generated with dompdf, and reports written after watching someone export to Excel and pivot it by hand. Finance teams are the harshest users you will have; build for them early.
React Native for iOS and Android against the same API as the web app, so a change to a rule does not have to be implemented three times. We ship and maintain both stores for Bright Future's e-learning platform in Kuwait, including video delivery over Bunny CDN.
If you already have something that half works, we read it before proposing to throw it away. The usual answer is to replace it module by module while the old system keeps running, rather than a big-bang rewrite that leaves you with nothing usable for a year.
Before we estimate anything, we spend time with the people who do the job today — what they type twice, where the process really breaks, which report the boss actually opens. Most of the requirements that matter are never in the brief.
We split the build into phases that each end in something you can put into real use, and quote them one at a time. You see working software early, and you are never committed to a year of unknowns on the strength of one document.
Weekly builds on a staging environment you can log into and poke at. Our two QA engineers test each release against the workflow we mapped rather than the wording of the ticket, which is how the gaps between the two get found.
We migrate the real data, run old and new side by side where the risk warrants it, and stay on afterwards. Nearly every platform we have launched is still being changed — that is normal for software built around a business, and we price for it rather than pretending it ends.
Send the process that keeps breaking — the file, the chat thread, the report someone rebuilds every Friday. We will come back with a straight view of what is worth building, what you should buy instead, and what a sensible first phase costs.