Knowledge Exodus
Deloitte Insights & eGain · June 2026

The $9 Trillion
Knowledge Exodus

Over the next four years, more than 30 million Americans will reach retirement age — triggering the largest single transfer of institutional knowledge in business history.

$9.6T
upper bound of projected
lost economic output
30M+
Americans reaching
retirement in 4 years
85%
of C-suite leaders see this as
moderate to mission-critical
92%
of organisations still fail to
capture retiring knowledge
THE 92% PROBLEM
The critical number
92%

of surveyed organisations still fail to consistently capture knowledge from soon-to-be retirees.

This is not a failure of intent

Eighty-five per cent of C-suite leaders already recognise this as a moderate to mission-critical threat. Boards are aware. HR programmes are running. Exit interviews are being scheduled.

The 92% statistic persists because it is a failure of method, not will. The tools organisations are using to capture knowledge are the wrong tools for the knowledge that matters most.

The tenure divide

The memory gap opening in real time

Average job tenure today
3.9
years
Down from 4.6 years a decade ago
The average worker is spending less time at each organisation — and accumulating shallower institutional memory at each stop.
Baby Boomer average tenure
8+
years
More than double the current average
This depth is not just time — it is context. The engineer who knows why a system was designed a certain way. The account leader who holds the history of a decade-long client relationship.

This knowledge will not naturally reaccumulate in the next generation of workers. The conditions that created it — long tenure, stable teams, repeated exposure to the same systems and clients over years — have structurally changed.

Why documentation fails

The knowledge that matters most
cannot be written down

Documentation-based capture — knowledge bases, exit interviews, wiki pages — structurally does not work for tacit, embodied, situational expertise.

Situational judgement

The practitioner doesn't know they know it until the moment they exercise it. It surfaces in action, not in answers to interview questions.

Relational context

Who to call when a decision stalls. The history behind a stakeholder's resistance. The unwritten rules of a market. None of this lives in a knowledge base.

Pattern recognition

Decades of exposure to how systems behave under stress, what precedes a failure, when to override the model. Invisible until it is gone.

"

The knowledge that matters most is precisely the knowledge that cannot be written down in a knowledge base, because the person holding it does not know they know it until the moment they exercise it.

Deloitte Insights & eGain · The $9 Trillion Knowledge Exodus · 2026
A structural diagnosis

Two failures hiding behind one demographic

01

The demographic failure

30 million Baby Boomers exiting the workforce over four years is an event. It is large, concentrated, and largely predictable. Organisations have had years to prepare — many have not.

The evidence: Average job tenure fell from 4.6 to 3.9 years even before peak retirement begins. The pipeline of deep institutional knowledge is thinning from both ends simultaneously.
02

The methodological failure

Even where organisations act, they reach for documentation. Exit interviews. Knowledge bases. SOPs. These tools capture explicit knowledge — the procedures and facts that can be articulated. They leave tacit knowledge untouched.

The consequence: 92% failure rates despite rising awareness. The tools and the target are mismatched at a structural level — not an execution level.
Where FSM comes in

Physical AI is not just a care technology. It is a knowledge technology.

The same AI systems now capable of supporting a disabled person through their daily routine are also capable of observing, encoding, and surfacing the situational expertise that documentation cannot capture. The embodied knowledge crisis in elder care — where knowledge leaves with carers, nurses, and social workers at every handover — is a version of the same structural problem the Deloitte/eGain report identifies at economy scale.

In care settings
Each shift handover loses context. AI systems that observe, not just document, can build a continuous picture of a person's needs, preferences, and patterns — without requiring the expert to narrate what they already know.
In institutional settings
The same principle applies to engineering teams, clinical departments, and policy units. Knowledge that lives in practice can be observed in practice — not extracted in an exit interview the week before retirement.
The FSM position
Physical AI that works for disabled people — that is responsive, contextual, and genuinely adaptive — is the same class of technology that can close the $9 trillion gap. People First is not a constraint on capability. It is the specification that makes it work.

Read the full Deloitte Insights report

The $9 Trillion Knowledge Exodus: How Organizations Can Turn Baby Boomer Retirements Into Competitive Advantage · Deloitte Insights & eGain · June 2026