This is a Field Note: a pattern I've observed repeatedly across organizations. It's a composite, not a single engagement, and I'm not claiming I solved it in any one place.
Symptom
Questions route to a person instead of a system. Why doesn't this invoice match the contract? Ask Susan. What did we promise this client in 2021? Ask Susan. How does the handoff actually work? Ask Susan.
What I Notice
There is rarely just one. Most organizations have a Susan per function: one for sales history, one for onboarding exceptions, one for billing anomalies. Each accumulated the role the same way — by being conscientious in the presence of a gap, repeatedly, for years.
Nobody designed this. That's what makes it hard to see: there was never a decision to point at.
What Is Usually Actually Happening
An undocumented database with excellent attendance. The employee is no longer performing a role; she is functioning as infrastructure. Her job description captures a fraction of her actual job, which means the org chart is describing a company that doesn't exist.
Critically, the Susans are evidence, not problems. Each one marks the precise location of a system that was never built. The workaround is doing a real job.
Why It Matters
Key-person risk is the obvious cost, and the one leadership names when she resigns. The quieter costs arrive earlier: processes can't be improved because nobody can see them, automation encodes the wrong thing because the real logic lives in someone's head, onboarding takes months because the curriculum is "sit near Susan," and Susan herself burns out doing two jobs, one of which is invisible and therefore unrewarded.
What I Would Investigate
Where do questions actually go, versus where the org chart says they go? Which decisions stall when one person is on vacation? Which processes have documentation that people bypass? And — before removing anything — what job is this workaround doing? Don't eliminate the workaround until you understand what it's compensating for, or you'll simply move the failure somewhere less visible.
Related Case Studies
The Build Wasn't the Problem. The Gap Was. — where I briefly became the Susan of a CRM implementation. Documented ≠ True — what happens when you ask five Susans the same question.