Skip to main content

New: Take the KSM™ AI Visibility Assessment and identify what is limiting your visibility in AI-generated answers. Start now

KSM Model — Knowledge Structuring Model official seal

Scoring standard

A transparent, versioned scoring methodology.

Every KSM™ assessment reports the methodology version, assessment date, evidence date, and limitations so results can be reviewed, reproduced, and challenged.

Composite score

How the 100-point score is composed.

AI Visibility = f(E × S × C) · E 40% · S 40% · C 20%

The Knowledge Structuring Model diagram showing the AI Visibility formula as a function of Extractable Structured Content, Entity Salience, and Citation Authority, with maturity stages and business impact.
The KSM™ model behind AEO and GEO visibility: AI Visibility = f(E × S × C), the three pillars, and how optimization evolved from SEO to AEO to GEO.
KSM™ pillar weights and scoring questions
PillarNameWeightScoring question
EExtractable Structured Content40%Can AI understand you?
SEntity Salience and Knowledge Graph Alignment40%Can AI recognize you?
CCitation Authority Reinforcement20%Can AI trust and cite you?

Pillar sub-scores are normalized to 0–100 and combined using the fixed weights above. Because the pillars interact multiplicatively in practice, the report always names the lowest-performing pillar as the constraint to address first.

Point allocation

How the 100 points are distributed.

Each pillar is assessed through named sub-criteria so a score can be traced back to observable evidence rather than judgment alone.

E · 40 pts

Extractable Structured Content

  • Schema validity — valid structured data that matches visible page content
  • Crawlability — retrievable, unblocked, server-rendered primary content
  • On-page hygiene — heading order, definitions, and canonical consistency
  • AEO structure — question-and-answer blocks an engine can lift directly

S · 40 pts

Entity Salience

  • Knowledge-graph presence — the entity resolves in public knowledge bases
  • Identifier consistency — names, IDs, and sameAs links agree everywhere
  • Cross-platform presence — consistent representation across major surfaces
  • Disambiguation — the entity is distinguishable from similarly named ones

C · 20 pts

Citation Authority

  • Peer-reviewed citation — scholarly or refereed references to the work
  • Third-party application — independent parties using or implementing it
  • Press citation — credible editorial coverage outside owned channels

Sub-criteria are scored from dated, observable evidence. The set of sub-criteria is fixed within a methodology version, which is what makes two assessments comparable.

Illustration

A hypothetical worked example.

The figures below are illustrative only. They describe an invented organization to show how a composite score is assembled — they are not a published assessment of any real organization, and no live or dated client scores appear on this site.

Hypothetical KSM™ score illustration for an example organization
PillarNormalized sub-scoreWeightWeighted contribution
E70 / 10040%28.0
S45 / 10040%18.0
C30 / 10020%6.0
Composite52 / 100

In this hypothetical case, citation authority is the lowest pillar, so independent corroboration would be sequenced first — improving schema coverage further would raise the strongest pillar while leaving the constraint in place.

Evidence

Evidence standards.

  • Observable and dated

    Each finding records what was observed, where it was observed, and the evidence date.

  • Authorized surfaces only

    Analysis is limited to the public website you confirm you are authorized to submit, plus publicly available third-party sources.

  • Schema must match visible content

    Structured data is credited only when it corresponds to content a person can see on the page.

  • Independent corroboration

    Citation authority credit requires sources outside your own domains and marketing channels.

Limitations

What a KSM™ score does not do.

  • It does not guarantee placement, citation, recommendation, traffic, or revenue in any AI system.
  • AI platforms, retrieval methods, data sources, and prompts change continuously; a score reflects the evidence available on the assessment date.
  • Coverage is limited to publicly observable signals. Private data, paid placements, and internal systems are out of scope.
  • Scores across organizations are comparable only when produced under the same methodology version.
  • Assessment output is diagnostic guidance, not legal, financial, or compliance advice.

Version history

KSM™ Methodology v1.0 (2026)

  1. v1.0 · 2026

    Initial published methodology aligned to the original SSRN working paper. Fixed pillar weights of 40 / 40 / 20 and a 100-point composite score.

Start the AssessmentRead the Original Research