Capturing Institutional Knowledge — Before It Walks Out the Door

Omar Farooq, Head of Analytics at the EdTech company in Karachi, learned about the institutional knowledge problem the hard way. When a Senior Project Manager named Layla resigned after eight years, the company sent around an email congratulating her on her next chapter and asking colleagues to sign her card. Her last day came and went. Three months later, a client called to say the new project manager had missed something important — a standing agreement that Layla had negotiated informally two years earlier, the kind of arrangement that existed in email threads nobody had filed and in a relationship that nobody else had the context to maintain.

The knowledge had not disappeared. It had evaporated. Slowly and invisibly, over the eight years Layla had been there, the organisation had loaded critical understanding into one person's memory — her read on the client's real decision-makers, her institutional sense of when to escalate and when to absorb, the unwritten rules she had learned by getting them wrong first. None of it was in any document. None of it survived her departure.

This lesson is about recognising when that process is happening and stopping it — not just when someone resigns, but before they ever think about leaving.

The Two Forms of Organisational Knowledge

Every organisation accumulates knowledge in two distinct forms. Understanding the difference is the starting point for everything that follows.

Form	Definition	Where it lives	At risk of permanent loss?
Explicit knowledge	Written, formalised, and codable	Documents, databases, policies, processes	Low — it can be found and transferred
Tacit knowledge	Unwritten understanding accumulated through experience	People's heads	High — it leaves when they leave

Explicit knowledge is findable, even if hard to find. The policy exists; the process is documented; the org chart is current. Tacit knowledge is different: it is the understanding of how things actually work that is not in any document and may never have been articulated.

Examples of tacit knowledge that organisations routinely lose:

• Which stakeholder's informal approval you need before the formal committee sign-off
• What the client actually cares about most (as opposed to what the contract says)
• Which internal process step always causes delays and why, and how experienced people navigate around it
• What the organisation tried in 2021 that failed and should not be tried again
• The three relationships that hold a delivery together, and what each one needs

When a long-tenured employee leaves, their explicit knowledge stays (it was never only theirs). Their tacit knowledge goes with them.

The Knowledge Risk Classification

Not all knowledge departure risk is equal. The /knowledge skill uses a five-factor scoring model to classify each knowledge holder as HIGH, MEDIUM, or LOW risk.

Five-Factor Scoring Model

Score each factor on a 1–3 scale, then sum for a total score:

Factor	1 — Low Risk	2 — Medium Risk	3 — High Risk
Tenure	Less than 2 years	2–5 years	5+ years
Role criticality	Support function	Specialist	Leadership or sole expert
Documentation level	Well documented	Partial documentation	Undocumented
Successor readiness	Successor fully prepared	Successor developing	No successor identified
Client / revenue impact	None	Some indirect	Direct and significant

Score interpretation:

• 5–7 (LOW): Standard handover document is sufficient
• 8–10 (MEDIUM): 2 capture sessions; 3–4 knowledge articles
• 11–15 (HIGH): Full capture programme; 3+ sessions; immediate escalation to leadership

Worked Example: David Okafor

David Okafor is Head of Client Services at a technology company in London. He joined 12 years ago as an account manager and has spent the last five years managing the firm's three largest clients — together, they represent 60% of recurring revenue. He is the company's primary relationship contact for all three. His knowledge of these relationships — the real decision-makers, the relationship history, the unwritten agreements — is held by almost no one else. He has just announced he is leaving in six weeks.

Risk scoring for David Okafor:

Factor	Score	Justification
Tenure (12 years)	3	5+ years: maximum tenure score
Role criticality (sole client relationship holder)	3	Leadership role; no one else manages these accounts
Documentation level (undocumented)	3	Client knowledge is relational; nothing written
Successor readiness (no successor named)	3	No named successor; interim coverage only
Client/revenue impact (60% of recurring revenue)	3	Direct and existential

Total: 15/15 — HIGH RISK

This is the highest possible risk score. Six weeks is not enough time to capture 12 years of client relationship knowledge fully. The correct response is immediate escalation to senior leadership and the maximum capture effort the timeline allows.

Proactive vs Reactive Capture

Mode	Trigger	Time available	Quality of capture	Priority
Reactive	Employee resignation confirmed	Notice period only	Lower — employee is mentally disengaging	Urgent but compromised
Proactive	Annual knowledge review for high-risk holders	Unlimited	Higher — employee is fully engaged	Systematic and thorough

Reactive capture is what most organisations do: they wait until someone announces they are leaving, then scramble. This produces worse outcomes than proactive capture for two reasons. First, the time available is fixed (whatever the notice period is) and usually insufficient for high-risk holders. Second, a departing employee's quality of engagement declines as their last day approaches — they are mentally transitioning, emotionally focused elsewhere, and often working longer hours as they try to hand over responsibilities.

Proactive capture — scheduled annually for employees who score HIGH risk — extracts knowledge while the holder is fully engaged, has time to think carefully, and is not emotionally attached to the outcome. The knowledge is richer, more honest, and more thoroughly verified.

A proactive capture programme looks like this:

• Annual knowledge audit: For every employee with a HIGH or MEDIUM risk score, schedule one knowledge capture session per year as part of normal HR operations
• Output ownership: The knowledge articles are reviewed and owned by the manager, not just filed somewhere
• Trigger for full capture: If a proactive review reveals knowledge is particularly concentrated and undocumented, escalate to a full capture programme immediately — do not wait for the resignation

The Three-Session Interview Structure

For HIGH-risk knowledge holders, /knowledge recommends three sessions of 60–90 minutes each. Each session focuses on a different knowledge domain:

Session	Focus	Key Questions
Session 1	Client and stakeholder relationships	Who are the real decision-makers? What is the relationship history? What are the unwritten rules?
Session 2	Methodology and process	Where does reality differ from documentation? What are the early warning signs? What decisions do you make vs escalate?
Session 3	Institutional context	What would you tell your replacement? What should we not try again? What are you most worried we'll get wrong?

The sequence is intentional. Session 1 starts with concrete, relational knowledge (easiest to articulate). Session 2 moves to procedural knowledge (requires more reflection). Session 3 ends with the deepest tacit knowledge — the institutional wisdom that often surprises even the employee themselves when they try to articulate it.

Sample Session 1 — Client Relationships

Using David Okafor as the subject, here is how Session 1 unfolds:

Opening (5 minutes):

"David, the goal of these sessions is to make sure the knowledge you have built up over 12 years here does not simply disappear when you leave. We want what is in your head, not what is in our documents. There are no wrong answers. This is not a performance assessment — it is institutional learning."

Core questions:

"Walk me through each of the three major clients. For each one: who are the real decision-makers — not just the named contacts, but the people whose opinion actually counts when a decision gets made?"

"What has gone wrong in each of these relationships and how was it recovered? What should the next person know about the relationship history?"

"What are the unwritten rules with each client? Things that are not in any contract but have become the real operating norms?"

"What does each client actually care about most? Not what the scope document says — what makes each one feel that the relationship is working?"

"Who else in our organisation has important relationships with these clients that the new account manager needs to know about and manage?"

Sample response from David:

"With Whitmore Group — they're the largest account — the named contact is their CTO, but the person you actually need to keep happy is their VP of Engineering, James. James doesn't approve anything directly, but if James is unhappy the CTO hears about it within 48 hours. We learned this in year two when we shipped something the CTO loved but James thought was the wrong approach. Six months of difficult conversations. Every time we present to Whitmore now, I send James a draft the week before so he can flag concerns before the formal meeting. That's not in any document."

This is institutional knowledge. It is specific, it is actionable, and it is completely inaccessible to anyone who was not in those conversations. After each session, /knowledge structures these responses into knowledge articles.

Using /knowledge — The Worked Example

With Omar Farooq coordinating, the EdTech company invokes /knowledge to generate a capture plan for David Okafor:

/knowledge

Employee: David Okafor, Head of Client Services
Tenure: 12 years
Departure date: In 6 weeks
Risk classification: HIGH (score: 15/15)

Critical knowledge areas:
1. Whitmore Group relationship — sole contact; unwritten operating norms
2. Foresight Capital relationship — 8-year history; complex stakeholder map
3. Delta Education relationship — recently troubled; recovery in progress
4. Delivery methodology — how actual delivery differs from documented process
5. Internal stakeholder map — who needs managing to deliver for these clients

Successor: Not yet named; interim coverage from senior account manager
Time available: 6 weeks notice period

Sample output:

TASK:          Knowledge Capture Plan — David Okafor
DOCUMENT TYPE: Knowledge Capture Plan
JURISDICTION:  UK
CONFIGURATION: Loaded: hr.local.md
SENSITIVITY:   CONFIDENTIAL

KNOWLEDGE CAPTURE PLAN: David Okafor
================================================================
Role: Head of Client Services | Tenure: 12 years | Risk: HIGH

KEY KNOWLEDGE AREAS (ranked by risk):
  1. Client relationships (Whitmore, Foresight, Delta) — sole holder;
     unwritten norms; revenue critical; undocumented
  2. Delivery methodology — reality vs documentation; internal stakeholders
  3. Institutional context — what has been tried; what should not be repeated;
     the next 12 months if David were staying

CAPTURE METHOD: Interview-based (3 sessions × 90 minutes)

SESSIONS:
  Session 1: Client Relationships — 90 min — Week 1
    Outputs: 3 × Client Relationship Profile (Whitmore, Foresight, Delta)

  Session 2: Delivery Methodology — 90 min — Week 2-3
    Outputs: Delivery Methodology Supplement (reality vs documentation)
             Internal Stakeholder Map

  Session 3: Institutional Context — 90 min — Week 4
    Outputs: Institutional Context Guide
             "What I Would Do in the Next 12 Months" (if staying)
             Final: Handover Summary for incoming account manager

INTERVIEW GUIDE — SESSION 1 (Client Relationships):
  Opening: [Goal framing — 5 minutes]
  Core questions:
    — Walk me through each client: who are the real decision-makers?
    — What has gone wrong in each relationship and how was it recovered?
    — What are the unwritten rules?
    — What does each client actually care about most?
    — Who else internally holds important relationships with these clients?
  Closing: Is there client knowledge we have not covered?
================================================================

Knowledge Article Format

After each session, the notes are structured into knowledge articles using this format:

KNOWLEDGE ARTICLE: Whitmore Group — Stakeholder Map and Operating Norms
================================================================
Author:    David Okafor | Captured by: HR Business Partner
Date:      [Date] | Reviewed by: VP Client Services | Status: DRAFT
Category:  Client / Stakeholder
Audience:  Successor Account Manager

THE KNOWLEDGE
Named contact is CTO David Whitmore. Real decision influence sits with
James Hargreaves, VP Engineering. James does not formally approve but
will surface concerns to David within 48 hours if unhappy with a decision.
Best practice: share all major proposals with James as a draft one week
before formal presentation to David. This allows James to raise technical
concerns before the formal meeting, preventing the 6-month difficult period
that followed the Q3 2022 delivery.

WHEN IT APPLIES
Every major proposal, roadmap change, or scope adjustment that goes to
CTO level. Also applies to incident communications — James gets a copy.

EXCEPTIONS AND EDGE CASES
Commercial/pricing conversations are CTO-only; James is not involved
and should not receive pricing information.

RELATED CONTACTS
David Whitmore (CTO): [contact details]
James Hargreaves (VP Engineering): [contact details]
Alice Park (Project Sponsor — approves budgets): [contact details]

RELATED DOCUMENTS
Whitmore Group — Contract and Scope: [location]
Whitmore Group — Account History 2018-2024: [location]

CONFIDENCE LEVEL
HIGH: This is how it works. David has operated this way for 5+ years.
================================================================

Your output will vary

The specific interview questions, capture plan structure, and knowledge article content will vary based on the employee's role, knowledge domains, and departure timeline. The teaching point is the process: risk assessment first, then session structure, then structured articles — not ad hoc documentation in the last week before departure.

Exercise: Knowledge Capture Sprint

Type: Institutional Memory Time: 45 minutes Plugin command: /knowledge Goal: Identify your highest-risk knowledge holder, generate a capture plan, and structure a sample knowledge article from a simulated interview

Step 1 — Identify Three Knowledge Holders at Risk

For your organisation (real or hypothetical), identify three employees whose departure would represent significant institutional knowledge loss. For each, score using the five-factor model:

Factor	Employee 1	Employee 2	Employee 3
Tenure (1–3)
Role criticality (1–3)
Documentation level (1–3)
Successor readiness (1–3)
Client/revenue impact (1–3)
Total
Classification

Select your highest-scoring employee for the rest of this exercise.

Step 2 — Generate a Knowledge Capture Plan

Invoke /knowledge with the details of your highest-risk employee:

/knowledge

Employee: [Name and role]
Tenure: [Years]
Departure timeline: [Resignation announced / no departure planned — proactive]
Critical knowledge areas: [List the top 3 knowledge areas at risk]
Successor: [Named / Not yet identified]
Time available: [Notice period, or for proactive: unlimited]
Jurisdiction: [UK / Pakistan / other]

Review the output for:

• Are the knowledge areas correctly prioritised by risk?
• Is the session structure appropriate for the risk level (HIGH = 3 sessions)?
• Do the interview questions go deep enough? Add follow-up questions where needed.

Step 3 — Conduct a Simulated Interview (20 minutes)

With a colleague (one plays the departing employee, one plays the interviewer): run a 20-minute simulated Session 1 using the interview guide from Step 2.

The interviewer's job: follow the questions, but pursue depth over coverage. One detailed answer about a real situation is more valuable than five shallow answers about hypothetical ones.

Note-taker format:

• What is the knowledge? (the factual content)
• What is the context? (when does this apply; why does it matter)
• What are the exceptions? (when does the standard approach NOT work)
• Who else should know this?

Step 4 — Structure the Captured Knowledge

Take your interview notes and invoke /knowledge to structure them into articles:

/knowledge

Please structure the following interview notes into knowledge articles.

Role context: [The departing employee's role and primary knowledge domain]
Target audience: [Successor / whole team / whole organisation]

Interview notes:
[Paste your notes from Step 3]

Verify each article includes: title, the knowledge, when it applies, exceptions, related contacts, and a confidence level (HIGH/MEDIUM/LOW).

Deliverable: A knowledge risk scoring table for three employees, a full /knowledge capture plan for the highest-risk employee, and at least one structured knowledge article from the simulated interview.

Try With AI

Try With AI

Use these prompts in Cowork or your preferred AI assistant.

Reproduce: Generate a knowledge capture plan for David Okafor.

I need to create a knowledge capture plan for a departing employee.

Employee: David Okafor, Head of Client Services
Tenure: 12 years at the company
Departure date: In 6 weeks
Risk level: HIGH

Critical knowledge areas:
1. Three major client relationships — he is the sole contact; no written
   record of stakeholder maps, relationship history, or unwritten norms
2. Delivery methodology — how actual delivery differs from what's documented
3. Internal stakeholder management — who needs managing to deliver for clients

Successor: Not yet named; senior account manager covering interim

Please produce:
1. A knowledge risk assessment with score justification
2. A three-session capture plan (session title, focus, key questions)
3. The knowledge article format I should use to document what comes
   out of each session

What you are learning: The capture plan — risk classification, session structure, interview questions, expected outputs — is the entire knowledge capture workflow. By generating and reviewing it for a concrete case, you learn what a complete plan looks like and can apply the same structure to your own organisation's highest-risk holders.

Adapt: Score your own organisation's knowledge risk.

I want to assess institutional knowledge risk at my organisation.
For each of the following employees, score them using a five-factor model
(each factor scored 1–3: LOW, MEDIUM, HIGH risk) and classify each as
overall HIGH, MEDIUM, or LOW risk.

Five factors: tenure, role criticality, documentation level, successor
readiness, client/revenue impact.

Employee 1: [Role, years of service, what they know, succession status]
Employee 2: [Role, years of service, what they know, succession status]
Employee 3: [Role, years of service, what they know, succession status]

For the highest-risk employee: what are the top three knowledge areas I
should capture first if they announced tomorrow that they were leaving?

What you are learning: The risk scoring model forces you to look at your organisation's knowledge landscape systematically. Most practitioners have never done this — they know intuitively that "we'd be in trouble if X left" but have never quantified it. Scoring multiple employees reveals patterns: where knowledge is dangerously concentrated, and where proactive capture is most urgent.

Apply: Design a proactive knowledge capture programme.

I want to design a proactive knowledge capture programme for my
organisation — one that does not wait for resignations to start.

Context:
- We have [N] employees total, approximately [X] with 5+ years tenure
- Our highest-risk knowledge domains are: [list 2-3 areas]
- We currently have no structured knowledge capture process

Please design a programme that includes:
1. Cadence: How often should we conduct proactive capture sessions?
   For which risk levels?
2. Format: What is the right capture method for different types
   of knowledge (relational vs procedural vs contextual)?
3. Output: Where should knowledge articles live? Who owns them?
   How are they kept current?
4. Prioritisation: How do we decide where to start — which employees
   and which knowledge areas first?

End with the single most important thing to get right in the first
90 days of running this programme.

What you are learning: The difference between a reactive and proactive knowledge capture programme is the difference between damage limitation and genuine institutional memory building. Designing the programme structure before a crisis forces you to think about cadence, ownership, and infrastructure — the things that determine whether knowledge articles get created and actually used, or created and filed away.

Flashcards Study Aid

---

Continue to Lesson 11: Offboarding and Knowledge Transfer →