A factory can install connected equipment, digital production systems, and advanced analytics, yet still rely on one surprisingly persistent dependency: the employee who understands how everything actually fits together. That person might export reports, fix coding errors, merge spreadsheets, or spot exceptions no one ever documented. The technology shows up clearly in investment plans, but the human work that keeps it all running often does not. And as Korean manufacturers continue to digitize, that gap is becoming more important.
Korea’s Manufacturing Data Still Passes Through Human Hands
A recent analysis by the Korea Small Business Institute, or KOSI, offers an unusually clear view of this gap. Using national data from the 2024 Smart Manufacturing Innovation Survey, the institute found that 92.4% of surveyed manufacturing SMEs collected manufacturing data, yet 75.7% still entered data manually.
The divide then continued after collection. While 52.1% analyzed manufacturing data, 99.7% of those analysis activities relied on internal personnel, and 72.7% of companies conducting analysis primarily did so themselves using tools such as Excel. KOSI also found that 99.2% of manufacturing SMEs surveyed did not have a dedicated organization for manufacturing data or AI technology.
Taken together, those figures reveal something more consequential than low software adoption. Manufacturing data may already exist inside a company, but employees still perform much of the practical work needed to make it usable.
That work can be easy to overlook because it often appears ordinary. Someone downloads a file, changes a label, reconciles two totals or explains why a figure should not be interpreted literally. Yet these small interventions can contain crucial business knowledge that the formal systems themselves do not carry.
Excel Is Often Evidence of a Dependency, Not the Problem Itself
Dohwan Kim, founder and CEO of Seoul-based manufacturing AI company Dfinite, sees spreadsheets differently from executives who regard them simply as outdated tools.
“Excel survives because it is the only place where a domain expert can join data across systems using their own judgment,”
Kim told ngopihangat in an exclusive interview on manufacturing AI challenges.
He described an experienced engineer exporting information held in separate company systems, placing it side by side and applying years of operational understanding to interpret the result. The spreadsheet is visible, but the employee’s reasoning about what to combine, ignore, correct or investigate may remain undocumented.
“That is genuinely valuable work. The problem is that it is invisible, unrepeatable, and leaves the company when the responsible engineer resigns,”
Kim said.
“Excel is not the disease; it is the symptom of missing decision infrastructure.”
This distinction matters because removing the spreadsheet does not automatically remove the dependency. A company can replace a file with another application while leaving the underlying interpretation concentrated in the same employee.
Manufacturing’s Hidden Integration Layer Can Be a Person
The pattern is not limited to Korean SMEs. The Manufacturing Leadership Council’s 2024 Data Mastery research found that 68% of surveyed manufacturers still used Microsoft Excel to analyze at least some manufacturing data.
Excel existed alongside more formal technology rather than instead of it. Corporate analytics systems were also used by 68% of respondents, statistical or business intelligence systems by 66%, and shop-floor analytics by 62%.
That coexistence changes the interpretation. Spreadsheets do not necessarily prove that factories failed to digitalize. They can reveal the points where digital systems still require human intervention before information becomes operationally useful.
The same survey helps identify who performs that work. Continuous-improvement teams were responsible for analyzing and generating manufacturing insights at 68% of respondents’ factories, followed by factory supervisors at 55% and operational-technology teams at 52%.
In other words, important analytical work often remains close to employees who understand production rather than sitting exclusively inside centralized technology departments.
Kim describes part of this burden as the hidden labor of skilled engineers acting as “human middleware.” Employees can spend time exporting, cleaning and reconciling information even though their formal role may be engineering, quality management or production supervision.
Effective Employees Can Make Organizational Gaps Harder to See
There is a paradox in this arrangement. A weak workaround creates a visible problem, but a highly capable employee can prevent the problem from appearing at all.
An experienced supervisor who knows how to reconcile contradictory reports keeps a meeting moving. An engineer who remembers how an unusual production condition affects a measurement prevents a false alarm. A quality specialist who knows which historical exception applies can resolve an issue before it becomes an escalation.
The organization experiences the successful result, not the invisible repair that produced it.
Over time, management can therefore mistake individual competence for institutional capability. A process appears reliable because the same employees repeatedly compensate for its gaps.
This becomes risky when nobody has mapped those interventions. The important question for executives is no longer simply which spreadsheets remain in use. It is which business processes depend on somebody knowing how to make those spreadsheets trustworthy.
Knowledge Transfer Competes With Daily Factory Work
Digitization does not automatically solve that dependency because employees also need time to transfer what they know.
A 2026 Korea Employment Information Service study examined digital transformation and changing skill requirements in the electrical and electronics manufacturing sector. Based on establishments with at least 10 employees, it estimated that 150,264 workers required training related to digital transformation, equivalent to 30.2% of the workforce covered by the study.
Only 30.7% of establishments supported training for existing employees. More revealingly, 41.6% identified insufficient training time caused by heavy workloads as the largest constraint on participation.
That creates another organizational contradiction. Employees whose experience is most valuable to colleagues can also be among the hardest people to remove temporarily from daily operations.
A plant may therefore recognize the need for knowledge transfer while continuously postponing it because production problems, customer commitments and routine responsibilities take priority. Expertise remains attached to individuals not because management deliberately designed it that way, but because daily execution leaves little room to convert experience into a shared organizational capability.
Documentation Cannot Capture Every Judgment
Manufacturers can respond by documenting procedures, formulas, reporting routines and common exceptions. That reduces key-person dependency, but documentation alone does not reproduce everything an experienced employee knows.
Operational expertise can include recognizing when the normal procedure does not fit the situation. It can involve knowing that an unusual reading is acceptable under one production condition, remembering why a workaround was created or understanding which colleague possesses context that never entered the official record.
This is one reason industrial knowledge transfer cannot be reduced to storing more files. The organization needs to identify where judgment enters a workflow and understand why an employee intervenes before deciding how much of that activity can be standardized or supported by technology.
Kim also points to knowledge retention as a value that can be difficult for manufacturers to quantify. When experienced operators retire, their tacit understanding of machine behavior, maintenance shortcuts, and process anomalies often leaves with them, creating gaps that are not immediately visible in production metrics but can affect long-term efficiency and troubleshooting.
“A retiring expert’s understanding of the plant becomes organizational infrastructure rather than a personal asset,”
Kim told ngopihangat.
And that distinction matters because losing an experienced employee can mean losing more than labor capacity. It can also remove the accumulated judgment that allowed everyday processes to function without anyone realizing how dependent they had become on one person.
Manufacturers Should Find the Invisible Work Before Automating It
This creates a practical mandate for manufacturers pursuing digital transformation and manufacturing AI. Before attempting to automate a workflow, companies should examine the unofficial work that allows the existing workflow to function.
Which spreadsheet would disrupt operations if its owner stopped maintaining it? Which management report requires manual correction before anyone trusts the number? Which recurring task depends on copying or reconciling information? Which exception is handled through experience instead of a written procedure?
Another useful question is even simpler: when the official information conflicts, who gets called?
These questions can expose parts of the organization that conventional system inventories miss. A software map may show applications and databases. An operational dependency map should also show the people whose judgment connects them in daily practice.
This matters to technology startups as well. A manufacturing product that replaces a manual step without understanding why employees created that step can remove useful flexibility along with inefficient work. The better opportunity may lie in distinguishing repetitive labor that should disappear from judgment that the organization needs to preserve.
The Spreadsheet May Be Disposable, but the Reasoning Is Not
Factories will continue adding automation, analytics and AI, and some spreadsheet-based work will inevitably disappear. That does not make the human layer irrelevant. It makes understanding that layer more urgent.
Korea’s own manufacturing data shows how much analysis still passes through internal employees, while workforce research shows that transferring new skills already competes with demanding day-to-day workloads. The challenge is therefore not simply getting workers to abandon familiar tools.
A spreadsheet can be replaced in an afternoon. Reconstructing the years of judgment that made it useful can take much longer.
Manufacturers that treat every manual workaround as technical debt may miss what those workarounds reveal. Hidden inside some of them is the operating knowledge that allowed the formal systems to function together in the first place.
The next stage of manufacturing digital transformation should not begin by assuming that invisible work has no value. It should begin by finding out where that work exists, why people perform it and what the company would lose if nobody remembered how.
Key Takeaway
- Korean manufacturing still depends heavily on human data work. KOSI found that 75.7% of surveyed manufacturing SMEs manually entered manufacturing data, while 72.7% of those conducting analysis primarily used tools such as Excel for their own analysis.
- Excel persists because it carries human judgment that formal systems don’t capture. It often becomes the place where experienced employees bridge gaps between disconnected data systems, making invisible decision-making part of the workflow.
- Skilled employees can function as “human middleware.” Their routine work can include reconciling, cleaning and interpreting information, even when those responsibilities are not visible in formal process maps.
- Successful workarounds can conceal structural weaknesses. Companies risk confusing individual competence with institutional capability when critical processes depend on knowledge concentrated in particular employees.
- Knowledge transfer faces a practical time constraint. Korea Employment Information Service research found that 41.6% of establishments studied cited heavy workloads and insufficient training time as the largest obstacle to vocational training participation.
- Manufacturing digital transformation should map invisible work before automating it. The global lesson for manufacturers and industrial technology startups is to identify the reasoning behind manual processes before removing the people, tools or workarounds that currently keep operations connected.
– Stay Ahead in Korea’s Startup Scene –
Get real-time insights, funding updates, and policy shifts shaping Korea’s innovation ecosystem.
➡️ Follow ngopihangat on LinkedIn, X (Twitter), Threads, Bluesky, Telegram, Facebook, and WhatsApp Channel.
🤝 Looking to connect with verified Korean companies building globally?
Explore curated company profiles and request direct introductions through beSUCCESS Connect.
PakarPBN
A Private Blog Network (PBN) is a collection of websites that are controlled by a single individual or organization and used primarily to build backlinks to a “money site” in order to influence its ranking in search engines such as Google. The core idea behind a PBN is based on the importance of backlinks in Google’s ranking algorithm. Since Google views backlinks as signals of authority and trust, some website owners attempt to artificially create these signals through a controlled network of sites.
In a typical PBN setup, the owner acquires expired or aged domains that already have existing authority, backlinks, and history. These domains are rebuilt with new content and hosted separately, often using different IP addresses, hosting providers, themes, and ownership details to make them appear unrelated. Within the content published on these sites, links are strategically placed that point to the main website the owner wants to rank higher. By doing this, the owner attempts to pass link equity (also known as “link juice”) from the PBN sites to the target website.
The purpose of a PBN is to give the impression that the target website is naturally earning links from multiple independent sources. If done effectively, this can temporarily improve keyword rankings, increase organic visibility, and drive more traffic from search results.
