Same Subsidy, Opposite Outcomes — The Variable Was the Vendor
It is not unusual for two plants that built systems under the same program track with comparable budgets to end up in opposite places. One sees equipment utilization improve within a year and cuts the time it takes to trace defect causes; the other stops opening the screens within six months. The track was the same, the funding was the same, and the feature lists in the proposals looked much alike. What separated them, in most cases, was who built it and who stayed with it.
Most adopting companies compare solution specifications. But a smart factory is not a product that ends at delivery — it is a relationship of at least three to five years covering maintenance and upgrades. If one developer leaves, the history behind the customizations disappears with them. If the vendor shuts down, there may be no way to extract the data you accumulated. What needs verification is not the solution, but the capability of the company that builds and sustains it.
Understanding the Supplier Capability Diagnosis
The Korea Smart Manufacturing Office operates a capability diagnosis program for supplier companies. Certified assessors evaluate technical capability, project management, quality management, and workforce operations against a standardized diagnosis model, and provide consulting on improvement directions by domain. The government covers 80% of the diagnosis cost, so the practical burden on suppliers is modest.
From 2026, a separate AI capability diagnosis is being operated. It assesses manufacturing-AI development capability — data preprocessing, model design and validation, and field deployment and operation — and assigns its own grade. Manufacturing-AI projects demand a different skill set from conventional MES or POP implementation, and the existing capability grade alone does not reliably indicate whether a vendor can deliver an AI project.
For Adopters: A Vendor Selection Checklist
Filter 1 — Registration and how current it is. Registration on the program management system is a minimum condition, not a qualification. Check whether the listed project record and staffing information have been updated recently.
Filter 2 — Diagnosis participation. Ask whether the vendor has undergone the capability diagnosis and the AI capability diagnosis, and if so, whether they will share the results. A refusal to share is itself a data point.
Filter 3 — Track record in your industry and process. Experience with the same process dramatically shortens the requirements definition phase. Where possible, visit a reference plant and ask the operators on the floor a single question: "Are you still using it?" A completed implementation and a system in daily use are two different things.
Four items must be confirmed in the contract.
For manufacturing-AI projects, add two clauses. Responsibility for securing training data (who collects and labels it, and in what volume) and the retraining cycle along with who pays for it. Without these, a drop in accuracy a few months into operation turns into a dispute over accountability.
For Suppliers: Why Take the Diagnosis
For a supplier, the capability diagnosis is closer to a health check from an outside perspective than to an examination. It surfaces weak areas that are hard to see from inside — absent configuration management, inconsistent project deliverables, dependence on one or two key engineers — and turns them into an improvement roadmap.
A manufacturing-AI grade functions as a trust signal at the proposal stage. Where capability previously had to be argued from past project lists alone, a third-party grade reduces the decision burden on the adopting company. Connecting the findings to hiring plans, quality management systems, and next year's program participation strategy turns the diagnosis from a one-time evaluation into a growth instrument.
Risk Management by Implementation Phase
How KITIM Can Help
KITIM supports vendor candidate verification, the design of proposal comparison scorecards, contract term review, and the establishment of phase-by-phase acceptance criteria. For supplier companies, we assist with preparation for the capability and AI capability diagnoses and with building improvement roadmaps from the findings. If you are preparing a smart factory project or find the vendor selection stage difficult to judge, please contact KITIM.