25 sites on a single-instance,
multi-site TrakSYS template.
What started in 2015 as a single TrakSYS implementation became the operational backbone of a global cork manufacturer — and the engagement that gave Kaizen Tech its name.
A global cork manufacturer with multi-site, multi-process complexity.
Corticeira Amorim is the world's largest cork producer, operating across multiple business units — natural cork stoppers, technical cork, composite cork, insulation, and floor & wall coverings. By 2015, operations spanned dozens of plants across Europe and beyond, each with its own legacy systems, paper routines, and locally-grown reporting.
The company didn't have a real-time visibility problem at any single plant. It had a standardization problem at the corporate level: comparing apples to apples across sites was nearly impossible, and rolling out an operational change required a parallel rollout in N flavors — the classic challenge of multi-site MOM deployment without a global template.
Lot genealogy and execution data trapped on paper.
Three structural issues sat at the core:
The second issue was governance. Each site had its own "MES" — usually a stack of Excel files, a custom database, and tribal knowledge. Comparing OEE across two plants required a translator. Pushing a new operating procedure took quarters, not weeks.
The third issue was the future the company knew was coming: AI in operations, predictive analytics, supply-chain optimization. None of that would work without a clean, contextualized, multi-site operational data backbone.
One template. One instance. Wave-based rollout.
We didn't propose 25 implementations. We proposed one — and a discipline for replicating it.
The architectural decision was a single-instance, multi-site TrakSYS template. One configuration, governed centrally, with explicit local-flex parameters at the site and line level. Sites couldn't fork the template — but they could configure within it. Anything else required a change-control process tied to corporate steering.
From the lighthouse, we built a phased rollout. Each wave grouped sites by similarity (geography, business unit, complexity) so the global template's stress points surfaced predictably. By wave three, the multi-site rollout was a repeatable program rather than a series of projects.
Architecture, integration, adoption.
- Master-data designEquipment, materials, BOMs, lot logic — modeled once, deployed N times.
- TrakSYS implementationCore build governed centrally; local configuration parameters explicit.
- OPC UA / MQTT connectivityDirect machine integration replacing manual data capture.
- ERP integration (SAP)Bidirectional ERP integration — production orders in, order confirmations, consumptions, and lot genealogy out.
- Paperless executionOperator tablets at the line. Quality, deviations, holds — all digital.
- Lot genealogyForward and backward traceability across all sites in a single query.
- Wave governanceSteering committee, change control, site-readiness gates.
- Hypercare to steady-stateEach site supported until the operations team could run it.
From paper to a corporate operational platform.
25+ sites are running on the shared instance today, with real-time visibility across the group. Paperless manufacturing execution replaced paper-driven processes at every site. Lot genealogy is queryable in seconds across the entire group. Corporate operating procedures roll out in weeks instead of quarters. Site-to-site OEE benchmarking is automatic, not artisan.
More importantly: the data backbone for AI-enabled operations exists. The next decade of Amorim's digital agenda — predictive quality, supply-chain optimization, autonomous routing — sits on the platform we delivered between 2015 and now.
All operational benchmarks referenced are illustrative based on prior engagements. Actual results require client-specific assessment.