Skip to main content

Knowledge Fog

Knowledge Fog is a way of measuring how unclear, incomplete, fragmented, or weakly grounded an organisation's knowledge is relative to its ontology and the evidence collected so far.

It is not only a metaphor. It should be treated as a practical metric, or family of metrics, that agents and people can use to decide where understanding is weak and where more discovery is needed.

Table of Contents​

Formal definition​

Knowledge Fog is the measured uncertainty and incompleteness of organisational understanding within a given scope, based on:

  • the current state of the ontology
  • the knowledge and evidence collected so far
  • the quality of relationships between known entities
  • the degree of contradiction, ambiguity, staleness, or missing ownership in the model

In simple terms:

Knowledge Fog measures how much of an organisation remains unclear, weakly structured, or insufficiently evidenced relative to the model it is trying to build.

What Knowledge Fog is for​

Knowledge Fog gives agents and organisations a way to say more than:

  • here is what we know

It lets them also say:

  • here is how complete our understanding is
  • here is where the model is thin
  • here is where hidden knowledge still dominates
  • here is where the agent should be cautious
  • here is where further discovery should happen next

Why it matters​

Without a concept like Knowledge Fog, agents can become overconfident.

They can treat partial knowledge as if it were complete, or produce recommendations without clearly distinguishing between:

  • well-modelled areas
  • weakly modelled areas
  • unknown areas
  • contested areas

Knowledge Fog gives the organisation a way to make uncertainty visible and actionable.

Core dimensions of Knowledge Fog​

Knowledge Fog should usually be expressed through multiple dimensions rather than one crude score.

1. Coverage Fog​

How much of the expected ontology scope is still missing.

Examples:

  • missing entities
  • missing domains
  • undefined capabilities
  • absent workflow objects

2. Relationship Fog​

How weakly connected the known entities are.

Examples:

  • missing links between workflows and roles
  • unconnected decisions
  • undefined dependencies
  • isolated knowledge objects

3. Evidence Fog​

How poorly grounded the current knowledge is.

Examples:

  • undocumented claims
  • low-confidence statements
  • no trace to source material
  • no linked artefacts or observations

4. Ownership Fog​

How unclear responsibility and stewardship are.

Examples:

  • no owner for a process
  • unclear reviewer for a decision
  • missing accountability for a capability or risk

5. State Fog​

How unclear or stale the current state of objects is.

Examples:

  • unknown workflow state
  • stale policy status
  • outdated system context
  • unresolved or drifting records

6. Consistency Fog​

How much contradiction or ambiguity exists in the model.

Examples:

  • competing definitions
  • conflicting evidence
  • inconsistent naming
  • multiple truths for the same object

7. Dependency Fog​

How much of the real dependency structure remains hidden.

Examples:

  • unknown upstream impacts
  • invisible downstream consumers
  • hidden cross-team reliance
  • undocumented system couplings

8. Recency Fog​

How much of the knowledge is old, stale, or weakly maintained.

Examples:

  • outdated process notes
  • no recent confirmation
  • old evidence with no review cycle

A simple scoring approach​

A first version does not need to be mathematically complex.

For a given scope, each dimension can be scored on a bounded scale, for example:

  • 0 = very clear / low fog
  • 1 = minor uncertainty
  • 2 = moderate uncertainty
  • 3 = significant uncertainty
  • 4 = severe uncertainty
  • 5 = very high fog / structurally weak understanding

This can be done at different levels:

  • per entity
  • per workflow
  • per domain
  • per team
  • per business capability

Example expression​

An agent might say:

  • Coverage Fog: 4
  • Relationship Fog: 3
  • Evidence Fog: 5
  • Ownership Fog: 2
  • State Fog: 3
  • Consistency Fog: 1
  • Dependency Fog: 4
  • Recency Fog: 2

That would tell the organisation that understanding exists, but it is still weak in coverage, evidence, and dependency structure.

What agents should do with Knowledge Fog​

A good organisational agent should use Knowledge Fog to:

  • decide where to ask questions next
  • determine when to be cautious in recommendations
  • identify where the ontology is underdeveloped
  • highlight areas dominated by hidden knowledge
  • prioritise discovery effort across domains
  • measure whether the KnowledgeFund is getting stronger over time

What a lower-fog state looks like​

An area has lower Knowledge Fog when:

  • its entities are represented in the ontology
  • relationships are well linked
  • evidence is traceable
  • owners are known
  • current state is visible
  • contradictions are limited and manageable
  • dependency paths are clearer
  • recent updates keep the model alive

What a higher-fog state looks like​

An area has higher Knowledge Fog when:

  • entities are missing or undefined
  • the ontology is weak or too abstract
  • most important knowledge is person-bound
  • links between concepts are absent
  • evidence is weak or missing
  • accountability is unclear
  • state is stale or unknown
  • dependencies are mostly hidden

Relation to KnowledgeFund​

KnowledgeFund provides the structure, ontology, and contribution system that allow Knowledge Fog to be measured meaningfully.

Without a KnowledgeFund, organisations usually still have fog. They just cannot see or discuss it clearly.