One platform across ERP and CRM.
A logistics group ran a fragile legacy bridge between ERP and CRM, with double entry on both sides. We replaced it with one observed integration layer — and the internal tools to match.
A bridge nobody dared to touch.
Orders were re-keyed between the ERP and the CRM because the sync script that was supposed to connect them had no owner, no logs, and a habit of failing silently.
Every change felt like a gamble, so the team routed around the system instead — six spreadsheet workarounds deep by the time we arrived.
- Double entry across ERP and CRM
- A legacy sync script with no owner and no logs
- Six spreadsheet workarounds covering the gaps
An integration layer with discipline.
We replaced the bridge with one API layer between the systems — built incrementally, flow by flow, with no big-bang rewrite and no downtime.
Testing, CI and monitoring went in from day one. Focused internal tools replaced the spreadsheets, and the documentation was handed to their own team.
- One API layer replacing the fragile bridge
- Monitoring and alerts on every flow — nothing fails silently
- Internal tools the team actually likes using
Stable, observed, and owned.
Double entry is gone, the layer has run reliably since launch, and their own team now ships changes to it — with tests, logs, and no gambling.
“Changes used to be a gamble. Now they’re a ticket.”
How the bridge got fragile.
The sync script was written years ago by a contractor who left, for systems that have both been upgraded since. It kept limping because nobody could afford to break it — the classic business-critical system nobody owns.
How the ten weeks ran.
Weeks 1–2 — Architecture review. We mapped every flow the old bridge carried, including the undocumented ones the spreadsheets had absorbed.
Weeks 3–8 — Build, flow by flow. Each integration shipped behind the old one, validated against live data, then switched over — no big bang.
Weeks 9–10 — Tools & handover. Internal tools replaced the spreadsheets; their team took ownership with docs, tests and runbooks.
A platform, not a patch.
On the operate retainer we watch the layer and ship the next tools as needs surface — warehouse scheduling is already in flight. New capabilities are weeks now, because the foundation is solid.