MES (Manufacturing Execution Systems) run production, collect data in real time and alert managers. Yet they overlook an essential resource: the field know-how of technicians.
For several decades, MES and other SaaS software have established themselves as the pillars of industrial management. These are valuable tools, certainly, but ones that too often remain distant from the realities of the shop floor. And it is precisely on the shop floor that performance is decided: in the technicians' actions, in empirical diagnostics, in the reflexes built up over successive interventions.
Mimorian is an industrial intelligence platform that models equipment, structures failure diagnosis and captures the know-how of maintenance teams through a multi-agent AI architecture. Its approach: build a functional digital twin of each piece of equipment and support technicians in their reasoning, so that field knowledge is structured naturally, with no extra effort.
So how do we bridge this gap between digital systems and operational reality?
Enrichment through the figures
Human error and unplanned downtime [Source: Vanson Bourne/ServiceMax, 2017]
In the manufacturing sector, 23% of unplanned downtime episodes are triggered by human error, compared with as little as 9% in the other sectors studied. A routine known only to the old hand, poorly passed on, ends up as machine downtime.
Growth of the MES market The global MES market is expected to grow from 16.3 billion USD in 2025 to 39 billion in 2034, an annual growth rate of 10.3%, which testifies to the scale of their adoption [Source: GM Insights, 2025]
The cost of poor knowledge sharing
Large companies lose an average of 47 million dollars per year because of ineffective knowledge sharing [Source: Panopto, 2018].
These figures tell us two things: management systems are here to stay, and poorly transmitted knowledge is expensive. Because beyond the figures:
- When an experienced technician leaves the company, a whole body of empirical knowledge often disappears with them.
- Breaks in the handover between teams are still common.
- The lack of knowledge capture directly limits quality, responsiveness and upskilling.
Why go beyond conventional tools?
ERP, MES, CMMS. Each plays its part in the digital factory. But between these building blocks, field technical knowledge remains the great forgotten asset. What technicians discover, adapt or resolve day to day is still passed on too often in an informal way, and ends up being lost like a form of industrial amnesia.
Yet this operational knowledge is valuable. It is what makes it possible to react quickly, to avoid repeated mistakes, to make progress from one failure to the next. Industrial intelligence platforms such as Mimorian set out precisely to bridge this gap. Rather than simply archiving, Mimorian models each piece of equipment in a relational graph (a functional digital twin), then orchestrates specialised AI agents to support the technician in their diagnostic reasoning. The report is generated automatically from the voice exchange, and each intervention enriches the factory's collective memory.
How do we structure what MES do not capture?
The solution lies in approaches that complement MES and ERP: tools able to structure field feedback, hands-on practices and empirical diagnostics directly on the shop floor. They still need to be grafted onto the systems already in place without rebuilding everything: our guide explains how to integrate a CMMS or an ERP into an existing information system.
This is the approach taken by industrial intelligence platforms such as Mimorian. The technician interacts by voice with an assistant that knows the machine, its components and its functional links. The AI proposes ranked hypotheses, guides them towards the relevant tests, and the technician confirms or rules out each lead. The report builds itself. The result: each technician becomes a driver of the collective memory, with no additional data-entry effort.
What does the knowledge MES never see look like?
Three examples, familiar in every plant:
- The restart setting. After a long stoppage, the line restarts with a parameter slightly different from the manufacturer's setpoint. The old hand knows it, the MES does not: the value lives in a notebook or in his head. It is one of the reasons why failures cluster at the restart after an annual shutdown.
- The early warning sign. A noise, a vibration, a temperature drift that precedes the failure by several days. The sensor reports a value; the technician recognises a pattern.
- The order of checks. On a known failure, the expert checks three points in a precise order and saves an hour. That order appears in no procedure.
Each of these practices is maintenance data in its own right. Structured as feedback reports linked to the equipment, they become searchable by the whole team; left informal, they leave with their holder.
Frequently asked questions
Can an MES capture technicians' know-how? The MES runs production: work orders, traceability, performance. Capturing maintenance reasoning calls for a complementary layer, one that structures the diagnosis and the feedback report at the time of the intervention.
How do you capture field knowledge without adding data entry? Through voice. The technician describes the intervention out loud; the AI transcribes and structures the report: symptoms, hypotheses, root cause, remedy. Capture happens during the work, with no extra form.
Where should you start? With a critical piece of equipment whose keys are held by one or two people. Structure the feedback reports on that scope, measure the diagnosis time saved, then extend.
Conclusion
Three points to remember.
- Field knowledge is industrial data, just like MES measurements: it can be structured and queried.
- The systems in place have a different role: ERP, MES and CMMS manage and trace; capturing hands-on practice calls for a dedicated layer.
- Capture succeeds when it lives in the flow of work: voice dictation during the intervention, automatic structuring, single data entry.
Next step: list, for one critical line, what only one or two technicians know how to do. That list is your risk map.
For an overview of the subject, read our complete guide: Capture your experts' know-how before they retire.
📚 Sources :
- Global Market Insights (GMInsights), 2025: Manufacturing Execution Systems Market Size & Share, 2034
- Panopto/YouGov, 2018: Workplace Knowledge and Productivity Report (press release)
- Vanson Bourne/ServiceMax, 2017: After the Fall, official press release (23% of unplanned downtime in manufacturing triggered by human error)