Michael Chang.
Let's talk →
Impact01Approach02Case Studies03Work Together04Insights05About Me06Let's talk →
Compliance · Middleware

e-Invoice
Middleware

Built from scratch in one month. Went live on a non-negotiable government deadline. 340,000 invoices later — still running.

Low-Code PlatformLHDN MyInvoissFTP IntegrationMulti-ERPRegulatoryAI-Assisted Dev
0+
Invoices processed
0+
Group entities on platform
0 month
Core development time
0 months
In production
Problem

A fixed government deadline.
No room to negotiate.

The Malaysian government mandated that entities above a defined revenue threshold implement e-Invoice — submitting structured invoice data to LHDN's MyInvois API for every transaction. The regulation was not optional. The deadline was not movable.

The group operates across 60+ legal entities, each running its own ERP system. The challenge was not just compliance for one system. It was compliance for an entire group, across multiple platforms, under a single deadline.

The finance and ERP teams spent months evaluating third-party middleware vendors — demos, RFPs, technical assessments. None met the combination of cost, scalability, and fit for a group-wide rollout. By end of April 2024, with three months to the deadline, the decision was made: build it in-house.

Solution

Why in-house won
over every vendor.

Months of vendor evaluation had already been exhausted. The commercial case for in-house was clear — a commercial middleware licensing fee multiplied across 60 entities, indefinitely, was materially higher than a one-time build. But cost alone was not the deciding factor.

Cost
60 entities. No per-entity licensing.
A commercial solution at scale across a group means licensing cost compounds with every entity added. An in-house platform scales to 60 entities — and beyond — on the same fixed development investment.
Scalability
One platform. All entities.
Group-wide rollout required a platform that could onboard every entity across the group without architectural rework. Custom-built meant the onboarding model was designed for this from the start — not retrofitted from a single-entity product.
Process Fit
Tailored to our invoicing process.
Commercial middleware is built for the average customer. The group's invoicing workflows, ERP landscape, and validation requirements are not average. An in-house build was shaped around the actual process — not the other way around.
Timeline Reality
3 months was enough — if we started immediately.
No time for further vendor negotiations. The internal team had proven low-code platform capability already demonstrated across multiple systems. The decision to build was also a decision to stop evaluating and start delivering.
Timeline
Jan – Apr 2024
3rd-party vendor evaluation — demos, RFPs, assessments. No viable option found.
End Apr 2024
Decision: build in-house on the existing low-code platform. Development starts immediately.
May – Jun 2024
Core development. AI-assisted coding. Intense ERP integration testing.
1 Aug 2024
Go-live. All 60+ entities. Compliance achieved on deadline.
Approach

Architecture & Build:
One month, sFTP, AI-assisted.

The key decision — use sFTP as the integration interface to all ERP systems, rather than building direct API connectors to each. Not every ERP in the group exposes a modern API. The group's various ERP systems each have different integration capabilities. sFTP is a universal protocol — every system can write a file, and every system can read one. By standardising on file-based exchange, a single middleware architecture served all ERPs without exception.

Unlock Full Case Study

This case study details proprietary architecture flows, integration strategies, and business outcomes. Please connect to find out more.

← PortfolioBack to portfolio
Next Case Study →Raw Material Dashboard