When the Process Is a Person
The hidden human infrastructure holding organizations together
I can almost-always tell who’s actually holding a place together, whether I’ve been there for years or I’m new on the payroll.
It’s rarely the person with the big-shot title. Yes, I’ve learned to spot this person almost on instinct now, because the pattern repeats itself so reliably that it stopped surprising me somewhere around the third organization I was working with.
Simply put, every organization has a person who knows how it actually works.
They know which form to use, where the current version lives, and why the instructions on the website are technically correct but practically useless.
They know that the request must go to Angela, even though the directory says Marcus handles it. They know which meeting needs to happen before the official meeting. And they know that the “final” document is not the file called Final_FINAL_v3_ACTUALFINAL, but the one buried nine replies deep in a thread whose subject line stopped describing its contents weeks ago. They know who can approve an exception, which deadline is flexible, and what happened the last time someone tried to follow the process exactly as written.
Sometimes this is their formal job, what they were hired to do. Often, it is not.
I’ve seen versions of this person in universities, corporations, nonprofits, volunteer groups, healthcare systems, and even families. Sometimes they are an administrator. Or a longtime employee, a project manager, a nurse, a receptionist, a longtime volunteer, or the colleague who has been around long enough to remember why everything is the way it is.
They are the person everyone asks for…well…everything. And because they usually know the answer, the organization appears to have one.
The organization’s memory has a name
We tend to imagine institutional knowledge as something an organization possesses. It lives in the handbook. The shared drive. The database. The onboarding materials. The carefully-named folder called “Resources.” But much of what allows an organization to function never makes it into those places.
It lives in people. It lives not only in what they remember, but also in the connections they make between things: the policy and the exception, the person and the history, the official workflow and the workaround that became necessary six years ago and, still is, to this day, never formally acknowledged.
A document can tell you what the process is supposed to be. A person can tell you what usually goes wrong: which step everyone skips, which form bounces back twice before it’s accepted, which approval actually needs two signatures instead of the one the policy lists. One is the process. The other is how the work actually gets done. And that difference matters.
Knowledge-management researchers have a name for that second thing, the how-it-actually-gets-done kind of knowing. They call it tacit knowledge, a term the chemist-turned-philosopher Michael Polanyi coined back in 1966 with a line that’s stuck around for sixty years because nobody’s said it better since: we know more than we can tell. Judgment, context, relationships, the small signals that tell an experienced person when the usual approach is about to fail—Polanyi was describing exactly this, decades before anyone called it a workplace problem. The American Productivity & Quality Center’s (APQC) more recent work on knowledge transfer keeps circling back to the same finding: so much of what an organization actually knows lives only in the heads of the people who’ve done the job long enough to know it.
Tacit knowledge is why a new employee can read every available document and still end up asking Angela for help.
Competence can hide a broken system
The person who knows how everything works is usually seen as extraordinarily capable. And they are.
But their competence can conceal just how much the organization depends on them. Each time they retrieve a missing file, explain an undocumented decision, redirect a request, or warn someone about a predictable complication, they slowly but surely close a gap in the system. And, because the gap gets closed, it rarely gets properly fixed.
The work continues. The deadline is met. The problem disappears before anyone with the authority to redesign the process has to see it. This creates a strange organizational illusion: the more effectively one person compensates for a weak system, the more functional the system appears. In a sense, the person becomes a bridge everyone mistakes for solid ground.
Eventually, questions begin to flow toward them automatically: Can you remind me why we do it this way? Do you know who has access to this? Where did we put that? Has this happened before? Who should I talk to?
No single question seems burdensome. Most take only a few minutes. But the person is no longer simply doing their own job. On a given day, they’re the department’s search engine when the intranet fails, its historian when nobody remembers why a rule exists, its translator when two teams use the same word to mean different things.
They become infrastructure. And infrastructure is easiest to ignore while it’s working. Nobody thinks about the plumbing until the pipe bursts.
I’ve noticed something else, too, watching this happen across enough organizations to stop calling it coincidence: it’s almost always a woman who ends up in this role. Often a woman of color. Not exclusively, but often enough that I no longer treat it as an accident. The unglamorous work of remembering, of keeping track of what nobody else bothered to write down, tends to land on the people already expected to absorb whatever the org chart doesn’t account for.
Researchers Linda Babcock, Brenda Peyser, Lise Vesterlund, and Laurie Weingart spent years studying exactly this pattern and gave it a name: non-promotable work, the tasks that keep an organization running and never show up on anyone’s review. Research by law professor Joan C. Williams and business researchers Katherine W. Phillips and Erika V. Hall goes further into why that work lands so disproportionately on women of color specifically, not just women in general: their studies on workplace bias found that women of color are more often expected to prove their competence over and over, and more often handed the office “housework” nobody’s tracking, the two expectations stacking instead of canceling out.
I didn’t need Linda and her coauthors or Joan and hers to notice the pattern; I needed them to confirm what I’d already watched happen, more than once, to the people I respected.
Knowing is work
We don’t always recognize memory as labor.
Remembering can look passive, as though information simply remains inside certain people because they happen to have good recall. But organizational memory is active work. Someone has to notice which details matter—and still remember, four years later, why they mattered enough to explain again and again. Someone has to know that two departments use the same word to mean different things, and catch the mistranslation before it becomes a problem.
Someone has to know not only where the information is, but also what it means.
This is especially easy to overlook when the person does that work generously. They answer quickly. They forward the old email. They explain the acronym again. They add the context that keeps everyone else from making the wrong call. Their knowledge becomes so reliably available that people stop seeing it as a contribution. It starts to feel like a feature of the organization.
Then the person goes on vacation. Or gets sick. Or retires. Or changes jobs. Or simply gets tired of being the human help desk. Suddenly, the organization discovers that a process used by twenty people was only ever understood by one. Their absence didn’t create the weakness. It simply revealed it.
Research keeps connecting employee turnover to organizational knowledge loss, and one thing that comes up often is timing: the loss can start well before someone actually leaves, especially in places with no real plan for capturing what that person knows. The risk isn’t only that a person will take information with them. It’s that the organization never valued them enough, while they were there, to notice everything they were holding together.
The handoff-document fantasy
Many organizations address this problem shortly before someone leaves.
They schedule a transition meeting. They create a spreadsheet. They ask the departing person to write down everything someone else might need to know, usually while that person is finishing projects, transferring responsibilities, training colleagues, cleaning out an inbox and a desk, and attempting to compress fifteen years of experience into a document called “Handoff Notes.”
The result may be useful. It’s rarely sufficient.
The problem is partly timing. Knowledge transfer can’t be treated as an emergency evacuation procedure.
But it’s also a problem of form. A checklist can preserve steps, but it can’t preserve judgment. A contact list can tell you whom to call, not why that person needs to be the one consulted first. The folder holds the final decision and loses everything around it—the rejected options, the competing pressures, the one small condition that changed the outcome. In my experience, in academic departments, the process-is-a-person person, is almost always the department secretary, the one person who knows which forms the dean’s office actually accepts, which committee assignments are political landmines, and why a policy exists that nobody else currently employed remembers writing.
Information storage is not the same as organizational memory.
Documentation is only the beginning
The solution to this problem isn’t simply to document everything.
Anyone who has searched a shared drive containing seven years of abandoned templates knows that more documentation can create its own kind of ignorance. The goal isn’t accumulation. It’s usable continuity, which means asking better questions while the knowledge is still alive and in use.
What decisions still depend on one person’s memory? Where does the written process diverge from the one people actually follow? And of what’s left—what needs a document, and what needs a mentor standing next to someone the first time they try a task and get it wrong?
Some of this belongs in a decision log, a process map, a clearly maintained system. Some of it only transfers through mentoring, shadowing, or the kind of conversation that happens because two people are in the same room on a Wednesday before the big meeting. And, frankly, some processes shouldn’t be preserved at all. Sometimes the person everyone depends on is carrying the knowledge of how to survive a process that should have been fixed years ago. In that case, the goal isn’t to document the workaround more beautifully. It’s to remove the need for it.
This is the larger design question: how do you build the structures that let knowledge move on its own, instead of relying on one person to carry it?
That question is only getting more urgent as organizations rush to point AI at their internal knowledge base. But even the research on human-AI decision-making keeps landing on the same limit: organizational knowledge is scattered across software, documents, and whatever’s still only in someone’s head. A system can search what’s been captured; it can’t reliably retrieve context that was never made visible in the first place. Put another way, technology might help you find the file. Someone still has to explain why that file matters.
The process should survive the person
The person who knows how everything works is incredibly valuable. The goal isn’t to make them less valuable, or to extract everything in their head before they can escape the organization. It’s to stop making their constant availability the price of keeping the place running.
Healthy organizations don’t eliminate human expertise. They build conditions where it can be shared, questioned, and carried forward. They notice when one person has become the route every uncertainty travels through, and they treat explaining things as real work, ongoing, not something crammed into the weeks before someone leaves. And they notice the other thing too: that being everyone’s only route in is exhausting, and that exhaustion is often part of why the person quits or retires at all.
Every organization has at least one bridge nobody built on purpose. It just showed up, over the years, made of goodwill and long memory and someone’s refusal to let things fall apart. I’ve spent my career learning to spot those bridges early. Increasingly, my job isn’t to walk across them. It’s to ask who’s building the second one, before the first one gets tired of holding everyone else up.