A procurement officer retires and takes forty years of supplier judgement out the door with her. A senior product manager leaves for a competitor and three years of design rationale vanishes. A customer-service team turns over twice in two years and nobody in the building can explain why the de-escalation script says what it says.
Every large organisation has survived this the same way for decades: replace the person with another person who, after a few painful years of apprenticeship, eventually learns the role well enough to keep it running.
AI is about to end that arrangement. And almost nobody is talking about what replaces it.
The tasks that juniors used to do are the tasks that juniors used to learn from. If AI does those tasks instead, the learning does not happen. The apprenticeship breaks. Ten years from now, there is no senior.
The problem is not where you think it is
This is not a training problem. It is a memory problem.
The junior buyer who processed hundreds of RFPs manually developed, over time, an intuition for which suppliers are spinning stories. AI does the screening now. The pattern recognition never builds.
The customer-service agent who took five hundred escalated calls in their first year learned — without being taught — when a complaint carries regulatory exposure, when it signals a retention risk, when it is the first warning that something is broken in product. AI summarises the call history now. The signal-recognition never builds.
The junior who drafted two hundred campaign briefs learned, through rejection after rejection, where the brand's voice boundary actually lives — not where the guidelines say it lives, but where it actually lives. AI drafts the briefs now. The voice-judgement never builds.
Same problem. Six different uniforms. And the answer cannot be "stop using AI" — the work is genuinely better, faster, cheaper, and no organisation is turning the copilots off.
The answer has to be: build the tuition somewhere else.
What aviation figured out fifty years ago
Commercial aviation solved a structurally identical problem decades ago. The sky is a shared, high-consequence, rapidly changing environment. No pilot can learn it through apprenticeship alone — the stakes are too high and the rate of change is faster than on-the-job experience can track.
So the industry solved the competence problem at three layers simultaneously.
Governance. Regulators — ICAO, EASA, the FAA — set the rules of the shared airspace.
Infrastructure. Airports, air traffic control, navigation aids, certified aircraft.
Individual competence, continuously maintained. Every commercial pilot spends recurring time in a simulator. Not as a one-off graduation event. As a recurring discipline. Type ratings renewed. Emergency procedures rehearsed. New aircraft types, new simulator programmes.
The simulator does something specific that is easy to miss. It preserves, in a testable form, the accumulated institutional judgement of the profession — every edge case, every emergency, every failure mode anyone has ever encountered — and it trains each individual pilot against that memory on a continuous basis. When a new failure mode is discovered, the simulator updates. Every pilot flying that type encounters the new scenario in their next check ride. The memory of the profession and the competence of the individual are coupled by design.
That is why aviation can replace pilots without losing the profession's knowledge. The knowledge lives in the simulator, not only in the pilots' heads.
The third layer, for the rest of us
Every knowledge-work function now needs exactly this third layer.
The first two are being built. The EU AI Act is binding. Sector-specific rules are coming. Procurement teams are writing vendor-assessment frameworks. Legal is writing data classifications for AI. IT is writing access controls. This is necessary work.
But the third layer — the equivalent of the simulator, the continuous maintenance of individual competence against a living memory of the role — is almost entirely missing.
A wiki is not a living memory. Neither is a policy document, a Notion page full of diagrams, or a quarterly training deck. Those are artifacts of codification — snapshots of what someone, at some moment, decided to write down. They do not update when the work changes. They do not test whether anyone absorbed them. They sit there, read by nobody, until an audit asks for evidence of governance.
A living memory is different. It contains scenarios, not documents — a thousand anonymised cases of actual supplier screenings, not a policy on supplier screening. It is measurable — does the buyer catch the authority-limit breach when the AI's confidence score is 97%? It updates from use — the edge case a practitioner catches today becomes a scenario for the next cohort tomorrow. And it trains proactively — when the measurement shows a specific person is weakening in a specific dimension, the memory generates targeted practice from real cases before the weakness produces an incident.
That is the third layer. Nobody is currently selling it.
What companies are actually doing
You can see the gap most clearly in who has noticed it.
The Big Four have noticed by contracting. KPMG cut its UK graduate intake by roughly 29% between 2023 and 2024. Deloitte cut 18%, EY 11%, PwC 6%. Across the sector, UK accountancy graduate adverts fell 44% against 2023 (City AM). The base of the pyramid is quietly shrinking. PwC's global chair has said explicitly that the firm wants "a different set of people." Pyramid becoming obelisk.
The big law firms have noticed by reconfiguring. Latham & Watkins flew all four hundred of its US first-years to Washington for an AI Academy and has repeated it with every cohort since. Ropes & Gray lets first-year associates spend up to four hundred non-billable hours a year on AI training, tool experimentation, and mentoring circles (Above the Law). Think about what that number means. The firm is paying, on the P&L, for competence formation that used to arrive as a by-product of billable work. It is buying the apprenticeship AI has broken.
IBM has noticed by expanding. In February 2026, IBM's Chief HR Officer Nickle LaMoreaux said: "We are tripling our entry-level hiring, and yes, that is for software developers and all these jobs we're being told AI can do." (Fortune) The bet is specific: if the rest of the industry thins its pipeline, there will be a senior-talent shortage in the mid-2030s, and the firms that preserved their juniors will own it.
And then there is everyone else. Most firms are deploying the tool that eliminates the first-draft work and leaving the apprenticeship question to the next CEO.
The capability gap ahead
Two or three years from now, organisations will have their governance in place. They will have their infrastructure in place. A regulator or a board will ask the obvious question — can your people recognise when the AI is wrong? — and they will reach for the evidence and find they have attendance logs and completion certificates.
Attendance is not competence. A certificate is a receipt. What the regulation will eventually want, and what the business genuinely needs, is demonstrated practical judgement, sustained over time, in a role-specific context.
That is not a product you can procure from a catalogue. It is a capability organisations will have to build, function by function — a living memory of the role, coupled to a measurement instrument, coupled to a proactive training loop. The three are one system. Any two without the third produce something that has been tried before and failed before.
The simulator is the point. Everyone else is building the airport.
A longer treatment of this argument — including how it applies inside procurement, customer service, product design, and marketing; what the four properties of a living role memory look like in practice; and why Glean, Cognitive Workforce Twins, and LMS platforms do not solve it — is in our research paper on the subject.
Sources
- City AM — Big Four slash graduate jobs as AI takes on entry-level work, June 2025. https://www.cityam.com/big-four-slash-graduate-jobs-as-ai-takes-on-entry-level-work/
- Above the Law — The Grace To Dabble: Two Biglaw Firms Look To An AI-First Future, November 2025. https://abovethelaw.com/2025/11/the-grace-to-dabble-two-biglaw-firms-look-to-an-ai-first-future/
- Fortune — Tech Giant IBM Tripling Gen Z Entry-Level Hiring According To CHRO, February 2026. https://fortune.com/2026/02/13/tech-giant-ibm-tripling-gen-z-entry-level-hiring-according-to-chro-rewriting-jobs-ai-era/