Why Every New Expert Repeats Some of Your Old Mistakes

A capable person joins the team. They read the handbook. They shadow for a week. They inherit a tidy folder of SOPs, tickets, and “lessons learned.” Three months later they ship a change, approve an exception, or run an appeal strategy that looks rational—and recreates a failure mode the organization already paid for once.

Nobody is confused about their credentials. The organization preserved the answers. It did not transfer the scars.

This is not a hiring problem

If the pattern only appeared with weak hires, the fix would be recruiting. It appears with strong hires. It appears when a solid engineer replaces another solid engineer. It appears when an operations lead rotates after a reorg.

It appears when a denial specialist inherits a book of work from someone who left with three weeks’ notice. It appears when a quality manager inherits CAPA folklore that sounds settled and is not. The common ingredient is not incompetence. It is replacement without inheritance of conditional experience.

Earlier pieces in this sequence named the underlying gap: documentation preserves state more reliably than experience; version history remembers what changed more often than why a change was believed safe; and the expensive knowledge—assumptions, rejected paths, scars, reconsideration conditions—still lives disproportionately in people’s heads. This piece is about what happens next. Every time that knowledge fails to travel with the role, a new expert pays tuition the company already paid.

What onboarding usually transfers

Most onboarding is honest about what it can do. It teaches the current procedure. It shows where the tools are.

It names the people to ask. It points at the approved answer for common cases. That is valuable.

It is also incomplete in a predictable way. The new expert learns what we do now. They rarely inherit:

• what we believed when we chose this path;
• which alternatives already failed, and for which non-obvious reasons;
• which workaround is load-bearing versus cargo cult;
• which exception was temporary and never got a retirement date;
• what evidence would make retiring the ugly rule responsible.

Those fragments are the difference between a procedure and experience. Without them, a smart newcomer does what smart people do. They reconstruct a plausible story from incomplete artifacts.

Sometimes the story is right. Often it is merely fluent. Fluent reconstruction is how old mistakes get re-bought under new ownership—quietly, with better slides, and with nobody present who can say, “we already learned that the expensive way.”

Replacement has many faces

Resignation is the loud case. It is not the only one. A team reorganizes and the decision moves from Product to Engineering to Operations. A vendor leaves and takes the real context with the contract.

A consultant finishes the engagement and the “knowledge transfer” session becomes a recording nobody revisits under pressure. A project that worked in one region is copied to another without the constraints that made it safe. AI adds another kind of new expert. An assistant or agent may know more about software, medicine, finance, or operations in general than any individual on the team.

It still was not present when your organization learned something the hard way. Treating retrieval over incomplete records as if it were tenure is the same failure mode with more polish. General intelligence is not organizational inheritance. A fluent narrative about why the system looks this way is not historical truth.

In a human handoff, someone who was there can still interrupt the wrong story. In an AI-assisted workflow, the wrong story can arrive at operating speed with the tone of settled expertise—and the next process change may encode it as fact.

“Give them more documentation” is not the fix

The tempting response is to thicken onboarding. Longer checklists. More mandatory reading.

Bigger succession packets. Interview the departing expert and store the transcript.

I am not against any of that when it is selective and honest. A clear handoff note beats folklore alone. A recorded walkthrough can save a team weeks. But endless documentation programs fail for familiar reasons.

People skip them under deadline pressure. They fill templates with ceremony. They write the approved answer and quietly omit the failed experiments that actually taught the organization something.

They produce artifacts to retrieve—not living experience with ownership, evidence, and a reconsideration condition. So yes—improve handoffs where they help. Do not confuse a documentation surge with converting tenure into something the next expert can actually inherit. If the strategy is “capture everything the last person knew,” the strategy will fail. Selective transfer—focused on high-consequence decisions and load-bearing exceptions—is more honest.

It is also still difficult. And it is not a substitute for keeping expert judgment in the loop. Practitioners should co-own what is worth transferring and when a reconsideration condition has actually fired. Extracting scars into a file so the expert can be sidelined is the wrong reading of selective inheritance.

Dogma is the other failure mode

There is an equal and opposite danger. Organizations can “transfer experience” by freezing it. “We tried that before.”

“That isn’t how we do things here.” “We standardized on this years ago.”

Past scars can prevent repeated mistakes. They can also prevent progress—especially when a new expert arrives with better tools, different constraints, or new evidence, and meets a permanent veto without conditions. The difference is whether inheritance includes the conditions around the lesson.

A poor transfer says:

We evaluated Approach X and rejected it. Don’t reopen it.

A useful transfer says:

We rejected Approach X because it could not meet requirement Y under constraint Z. Reconsider if Y or Z changes materially—or if new evidence contradicts the original failure mode.

The first turns onboarding into permanent policy by accident. The second transfers experience. Good institutional inheritance should make lessons easier to reuse and easier to challenge intelligently. Handing a new expert a list of forbidden ideas without reconsideration conditions is not memory.

It is dogma with a buddy system.

What selective experience transfer looks like in principle

For high-consequence work—not every habit, not every ticket—the useful inheritance is familiar by now, as diagnostic probes rather than a schema for a new platform, and not a shopping list that replaces ADRs, SOPs, CAPA records, lessons-learned notes, or operating reviews you already run. What was happening? What problem, incident, constraint, or observation triggered the judgment the newcomer is about to touch?

What did we believe then? Not what we know after three more failures.

What evidence was available at the time? What assumptions did the decision depend on? What alternatives were rejected, and why?

What did we expect to happen? What actually happened?

When should we reconsider it? What change in regulation, technology, workload, vendor behavior, evidence, or objectives would make the old lesson the wrong guide?

Picture—illustratively, not as a measured field study—a replacement engineer facing a brittle payment path. Git and tickets can show the diff and the incident. Useful inheritance also carries the assumption about provider idempotency, the cleaner alternative that made production worse, and the evidence that would make deleting the ugly backoff safe.

Or a new denial specialist inheriting appeal playbooks. The procedure lists documents. Experience transfer also names which wording used to work, which exception tends to apply to a patient class, which strategy still sounds reasonable while consistently failing, and which payer-policy or contract change would reopen a closed path—without pretending to replace the EHR and billing tools already in place.

Or an ops lead inheriting a “temporary” exception. The runbook states the exception. Useful inheritance states why it exists, who owns it, what risk it contains, and what condition retires it. In principle form—not product form—useful selective transfer still asks whether a lesson has a named owner, a link to the evidence that justified it, and an explicit retirement or reconsideration trigger.

Those are diagnostic checks over existing artifacts, not purchase criteria for a platform. Without those three, “knowledge transfer” is still folklore with better filing. Someone still has to decide what is worth preserving for the next expert and what stays out of scope—ideally people closest to the work, not only a central documentation mandate.

People remain accountable for the decisions. The organization would only retain more of the evidence and conditions that should travel with the role.

Paying tuition twice is a choice

Every established company has scars it would rather not repurchase. Some repurchase is inevitable. People are not tape recorders.

Contexts change. Not every lesson deserves permanent institutionalization. But systematically hoping that every new expert—or every new AI agent—will rediscover the hard parts by living them again is not a rite of passage.

It is an operating strategy with a predictable invoice. The invoice shows up as repeated incidents, revived initiatives under new branding, exceptions that never retire, and fluent reconstructions that harden into process. Onboarding will not fix that by getting longer.

Dogma will not fix that by getting louder. Selective institutional experience transfer—scars, assumptions, and reconsideration conditions for the work that actually matters—is the honest alternative to both.

The open operating question

New experts will keep repeating some of your old mistakes for as long as the organization preserves answers without the conditions that made those answers trustworthy. That is not a failure of talent. It is a failure of inheritance.

Knowledge bases help. Succession interviews help.

Better commits and clearer ADRs help. None of them automatically make the next person—or the next agent—different from the last when the unusual case arrives. The hard question is no longer only whether expensive knowledge is trapped in someone’s head. It is whether, at replacement, the organization can transfer enough why that the new expert does not have to invent a plausible story.

If scars and reconsideration conditions never travel with the role, the company will keep paying tuition it already paid. What would it mean for an organization to remember why—selectively, with ownership and challenge conditions—so the next expert inherits experience rather than rediscovering the hard way?