Most Manufacturing Digital Maturity Models Are Theory. This One Is Designed for 110 Plants

A manufacturing digital maturity model is only useful if a plant manager can find themselves in it. Most cannot. Ask an AI assistant for one and you get five neat levels that describe no factory anyone has ever worked in.

Roman Lafage built a three-level version instead, designed to work across more than 110 ofi manufacturing facilities.

Why This Manufacturing Digital Maturity Model Has Three Levels, Not Five

Lafage leads the Global Digital Operations Center of Excellence at ofi, the olam food ingredients business. Notably, his team resisted the urge to specify everything.

The model runs Digital, then Integrated, then Autonomous. Digital means removing paper and spreadsheets. Integrated means connecting MES, connected worker, and historian data into one visualization and reporting layer. Finally, autonomous means the system learns and optimizes itself.

Specifically, the goal was a model simple enough that most plants could locate themselves on it. Detailed enough to be correct is a different objective, and a worse one.

“The reality of many manufacturing sites in the world today is that there is still a lot of Excel, there is still a lot of manual work, paper, used all along on the shop floor.”

Roman Lafage, Vice President of Digital Operations, ofi

The Assessment Runs Bottom-Up

Most plant managers know the other version of this. Corporate picks your site, hands you a project, and the deadline is Q2 because global said so.

Consequently, ofi inverted it. Specifically, the central team interviews the plant head, maintenance, quality, and production before arriving. They spend the week on site refining the problems, then pick from the catalog against those problems.

“We are not coming and saying: you need that. We are more coming and saying: okay, what are the problems? What are the areas where digital can bring value?”

Roman Lafage, Vice President of Digital Operations, ofi

Furthermore, the week is educational in both directions. The plant learns what digital can do. In turn, the central team learns what the plant actually needs.

Four Value Pillars, Running in Parallel

ofi measures digital against four pillars: manufacturing efficiency, product quality and safety, asset optimization, and integrated supply chain.

Importantly, these are swim lanes rather than stages. For example, a plant in high season with downtime problems focuses on asset optimization. By contrast, a plant drowning in paper focuses on manufacturing efficiency. Ultimately, the site’s reality picks the lane.

The Catalog Is the Scaling Mechanism

Additionally, ofi built a shared digital catalog to stop plants across the network from solving the same problem separately.

It has three layers plus a floor. The backbone carries MES and LIMS. The applications layer carries faster wins: OEE, connected worker, energy monitoring. Similarly, the data layer carries process analytics and historians. Underneath sits foundations: ERP practice, the data platform, master data, control systems, infrastructure, and cybersecurity.

Therefore the foundations decide everything above them. Lafage is direct about it: an agent working at ERP level fails if the ERP data is not clean.

Technology Exposes Problems. It Does Not Solve Them

This is the most useful thing in the episode.

“It’s not because you bring a technology that you resolve all your problem. Usually that will make all the problem visible, that on the process side, on the capability gaps you have.”

Roman Lafage, Vice President of Digital Operations, ofi

In other words, the software does not create the gap. Instead, it removes your ability to ignore it. As a result, the work becomes organizational rather than something the IT department delivers.

The Entry Point Is No Longer a Nine-Month Project

Moreover, the cost of starting has collapsed.

Connected worker platforms deploy in weeks. In addition, clamp-on sensors give a line OEE visibility without a rebuild. Likewise, energy monitoring no longer requires a capital project.

“Digital, this first step is not inaccessible. I think there is very simple way now to get there.”

Roman Lafage, Vice President of Digital Operations, ofi

What the Plant Team Looks Like Next

Ultimately, the capability question outlasts the technology question.

For instance, Lafage took a call from plant process engineers asking for their own access to AI and data development tools. Three years ago that request would not have existed. In fact, he treats it as the clearest signal of how quickly the capability profile of a plant is changing.

Asked what he screens for now, his answer was one word: adaptability. Notably, he calls the people who have it super adapters, and he thinks the companies that build teams of them will pull ahead.

Frequently Asked Questions

What is a manufacturing digital maturity model?
A manufacturing digital maturity model maps the stages a plant moves through as it digitizes, from paper and spreadsheets to connected systems and finally self-optimizing operations. Specifically, ofi uses three levels: Digital, Integrated, and Autonomous. Above all, the point is that a plant manager can locate their own site on it.

How many levels should a manufacturing digital maturity model have?
Fewer than most vendors suggest. ofi settled on three after concluding that a manufacturing digital maturity model only works if plants can place themselves on it quickly. Admittedly, five-level models carry more detail, but detail is not the goal. In addition, plants can add granularity locally where a domain demands it.

What is the first step in manufacturing digital transformation?
Removing paper and Excel from the shop floor. Notably, Roman Lafage calls this level one, and he notes most plants worldwide are still there. Specifically, connected worker platforms digitize standard operating procedures and checklists, which creates the first reliable data stream everything later depends on.

Why do digital transformation pilots fail to scale across plants?
Usually because each site solves the same problem separately, and because the foundations underneath are weak. Consequently, ofi addresses the first with a shared digital catalog. Similarly, it addresses the second by treating ERP practice, master data, and infrastructure as prerequisites rather than afterthoughts.

What are the foundations required to scale AI in manufacturing?
ERP practice, a data platform, master data management, control systems, infrastructure, and cybersecurity. Indeed, Lafage is blunt about the consequence: an agent operating at ERP level will fail if the ERP data is not clean. However, foundations can progress in parallel with quick wins.

What is a digital catalog in manufacturing?
A centrally managed set of vetted solutions that plants select from, paired with a repeatable deployment method. Specifically, ofi structures its catalog in three layers: a backbone of MES and LIMS, an applications layer for faster wins, and a data layer for analytics.

How long does it take to digitize a manufacturing plant?
Far less than it once did. For example, connected worker platforms deploy in weeks, clamp-on sensors deliver OEE visibility without a rebuild, and energy monitoring no longer needs a capital project. In short, the nine-month implementation is no longer the entry point.

What skills do manufacturing plant teams need for digital operations?
Adaptability, first. Indeed, Lafage screens for it directly and calls the people who have it super adapters. In addition, plant leadership needs enough data literacy to discuss technology credibly. Meanwhile, process engineers increasingly request AI tooling themselves rather than waiting on central IT.

Bottom Line

The plants that scale are not the ones with the best technology. Ultimately, they are the ones ready to act on what the technology shows them.

To submit a request for a new episode topic from Practitioners Unplugged,
visit our contact page.

Explore more about Schneider Electric & AVEVA

Sponsored by Optimum Productions