The Answer Exists. It's Just in Someone's Head.

The binder behind the med room door has a corner folded over. Inside, taped to the inside cover, is a Post-it with three lines: 'If the barcode scanner beeps twice, hold the trigger for three seconds. Then say "reset" out loud. Don’t touch the power button.' No one remembers who wrote it. Everyone knows not to delete it. It’s not in Epic. It’s not in the shared drive. It’s there because the last person who fixed the scanner didn’t trust the manual — and no one else had ever seen it done.

We call this a gap in knowledge. We say the hospital needs better documentation. We buy new platforms, assign content owners, run compliance trainings. But the problem isn’t that the answer isn’t written down. It’s that the person who knows it isn’t the one who wrote it. And when they’re off, no one knows who to ask.

The answer isn’t missing. It’s invisible. Not because it doesn’t exist — but because we’ve designed systems to reward documents over people. We assume knowledge lives in files. It doesn’t. It lives in the rhythm of a shift, in the way the unit secretary pauses before answering the phone, in the glance between the charge nurse and the pharmacy tech when a new order comes in.

The application analyst spent months building a searchable knowledge base. Every workflow was mapped. Every policy was scanned. Every SOP was tagged with role, location, and version number. The leadership dashboard showed near-total adoption. The training was mandatory. The completion rate was high.

And yet, the new hire still walked over to the charge nurse every time the Pyxis gave an error. Not because they couldn’t find the document — they’d read it twice. But because the document didn’t say what the charge nurse had said last week: 'It’s not the machine. It’s the med cart battery. Pull the drawer out and wiggle the cord. Then try again.'

The document was correct. The context wasn’t. And the system had no field for ‘what happens when the manual doesn’t match reality.’

We keep designing for the person who writes the answer. We forget the person who needs to find it — and who needs to find it fast, under pressure, without reading a manual. The new hire doesn’t need more documents. They need a name. A shift. A pattern. A way to know who’s seen this before.

It’s not that the knowledge is absent. It’s that the system treats it as if it were. We treat every unknown like a blank page that needs filling, when in fact, it’s a voice that needs tuning. The charge nurse doesn’t need a new policy when the IV pump firmware updates. They just need to tell the next person: 'Now it’s the blue button. Don’t touch the red one anymore.'

We don’t train people to find people. We train them to find files. And so when the file doesn’t match what they’ve seen happen, they’re stuck. They don’t know who to ask. They don’t know how to ask. They don’t know if the person who knows is even still on shift.

That’s real, and I won’t wave it off. Some of these notes are wrong. I’ve seen them: 'Give d50 for hypoglycemia even if the patient is NPO' — written by someone who didn’t know the patient had diabetic ketoacidosis. If we just digitize every sticky note, we’re not fixing the problem. We’re preserving the risk.

But that doesn’t mean we should stop looking for the answers that live in people’s heads. It means we need to separate the dangerous from the useful. The dangerous knowledge needs to be surfaced, challenged, and corrected. The useful knowledge — the one that saves time, prevents errors, reduces frustration — needs to be connected to the person who holds it.

Solution Compass exists to handle the retrieval problem. Not by storing more documents, but by letting the new hire ask: 'Who fixed the printer on 3 East last week?' and getting a name, a shift, a pattern — not a PDF. The answer isn’t in the file. It’s in the person who did it. And that person is still here.

The preceptor doesn’t teach the new hire how to use the system. They teach them who to ask. 'If the order set won’t load, find Carlos on shift 2. He’s the one who rebuilt it after the last update.' That’s not training. That’s tribal knowledge. And it’s the only thing that works when the system fails.

We’ve made it so hard to find a person that we’ve trained people to find a document instead. But documents don’t adapt. People do. The charge nurse doesn’t need a new policy when the IV pump firmware updates. They just need to tell the next person: 'Now it’s the blue button. Don’t touch the red one anymore.'

We treat knowledge like a static asset. It’s not. It’s a conversation. And the best version of that conversation lives in the room, not the repository.

The binder isn’t full because we lack documentation. It’s full because we haven’t built a way to say: 'Ask Janice. She knows.'

I’ve watched the same thing happen on three different units in three different hospitals. The same note. The same printer. The same Pyxis glitch. The same person who fixed it last time — now on leave. The same new hire, standing there, holding the manual, knowing it’s wrong but not knowing who to call.

We don’t have a knowledge problem. We have a connection problem. The answer exists. It’s just in someone’s head. And when that person leaves the room, we act like the answer vanished.

We keep buying authoring tools when we need retrieval tools. We keep building repositories when we need relationships. We keep training people to search — instead of teaching them to ask.

The next time someone asks for a protocol and you point them to the SharePoint folder, pause. Ask yourself: Is this the answer they need? Or is it the answer someone wrote after the person who actually knew was on vacation? The real knowledge never left. It just stopped being visible.

The binder still sits there. The Post-it still sticks. And the next person who walks in will find it — because no one else has the courage to throw it away. They don’t know who wrote it. But they know someone did. And that’s enough.