AIRepScore™ Methodology Disclaimer & Interpretation Guide
Version 1.0Effective Date: Bussum, 15/06/2026
1. Purpose of this Document
This document explains how AIRepScore™ scores, analyses and reports should be interpreted.AIRepScore is designed to measure how artificial intelligence systems perceive, position, compare and recommend brands, organisations, products and services.
AIRepScore does not measure objective truth, market reality, customer satisfaction or business performance.
The purpose of this document is to provide guidance on the interpretation and appropriate use of AIRepScore results.
2. What AIRepScore Measures
AIRepScore measures AI perception.AI perception refers to the way large language models (LLMs) interpret, structure and associate information about a brand based on:
• Training data
• Learned associations
• Public information
• Retrieved web information
• Contextual prompts
AIRepScore evaluates four analytical dimensions:
AI Credibility
How credible, trustworthy and recommendation-worthy a brand appears to AI systems.
AI Brand Perception
How AI systems position and describe a brand.
AI Positioning
How AI systems compare a brand against alternatives and competitors.
AI Recommendation
How likely AI systems are to recommend a brand within a specific use case or decision context.
3. What AIRepScore Does Not Measure
AIRepScore does not measure:• Objective company quality
• Product quality
• Customer satisfaction
• Market share
• Revenue
• Brand awareness
• SEO performance
• Advertising effectiveness
• Financial strength
• Legal compliance
• Future business success
A high AIRepScore does not guarantee business success.
A low AIRepScore does not imply business failure.
4. AI Perception Is Not Objective Truth
AI systems do not possess direct knowledge of reality.AI systems generate outputs based on patterns, associations and information learned during training and supplemented by retrieved information.
As a result:
• AI systems may be incomplete
• AI systems may be outdated
• AI systems may contain biases
• AI systems may contain inaccuracies
• AI systems may disagree with one another
AIRepScore measures those perceptions.
It does not validate whether those perceptions are factually correct.
5. The Three Layers of AI
AIRepScore is based on the observation that modern AI systems operate through three distinct layers.Layer 1 – Application Layer
The application layer represents the interface between the user and the AI system.
Examples include:
• ChatGPT
• Microsoft Copilot
• Gemini
• Perplexity
• Claude
This layer may include:
• User history
• Personalisation
• Geographic context
• Time-based context
• Application-specific instructions
Different users may therefore receive different answers from the same underlying model.
Layer 2 – Trained AI Layer
The trained AI layer contains the knowledge and associations learned during model training.This layer forms the foundation of AI perception.
It contains:
• Brand associations
• Industry knowledge
• Market positioning
• Historical information
• Learned patterns
This knowledge remains relatively stable until a new model version is released.
AIRepScore refers to this as trained perception.
Layer 3 – Web Retrieval Layer
Many AI systems perform web searches before generating answers.The retrieved information may reinforce, modify or contradict the model’s existing understanding.
AIRepScore measures this influence through its Web Search Correction (WSC) methodology.
This layer represents the dynamic component of AI perception.
6. Why Scores Change Over Time
AIRepScore scores are not static.Scores may change due to:
• New AI model releases
• New training data
• Web search results
• News events
• Reputation events
• Industry developments
• Competitor activity
• Changes in online content
A score change does not necessarily indicate a change in company performance.
It may indicate a change in AI perception.
7. Understanding AI Credibility
AI Credibility measures whether AI considers a brand sufficiently credible and trustworthy to recommend.The AI Credibility Score consists of six dimensions:
• Authority
• Brand Clarity
• Trust
• Strategic Relevance
• Differentiation
• Sentiment
These dimensions evaluate the foundational conditions required before recommendation can occur.
A brand may have a strong market position but still receive a lower AI Credibility score if AI systems struggle to clearly understand, trust or differentiate the brand.
8. Understanding AI Brand Perception
AI Brand Perception measures what AI believes a brand stands for.This differs fundamentally from AI Credibility.
AI Credibility asks:
Can AI recommend this brand?
AI Brand Perception asks:
For what reason would AI recommend this brand?
Brand Perception evaluates how AI associates a brand with characteristics such as:
• Innovation
• Sustainability
• Expertise
• Reliability
• Customer Centricity
• Premium Positioning
• Affordability
• Emotional Appeal
These associations are typically more stable and evolve gradually over time.
9. Authority versus Expertise
Authority and Expertise are intentionally measured separately.Authority
Authority measures whether AI considers a brand a recognised and credible reference point within its category.
Authority relates to influence, reputation and category leadership.
Expertise
Expertise measures whether AI associates a brand with specialist knowledge, know-how, thought leadership or technical competence.
A brand may have strong expertise without being considered a category leader.
Similarly, a category leader may have strong authority while not being strongly associated with specialist expertise.
10. Understanding AI Recommendation
AI Recommendation is not a prediction of future customer behaviour. It is a measurement of recommendation likelihood within AI systems.Recommendation scores indicate:
• Whether a brand enters the AI consideration set
• How frequently it is suggested
• How highly it is ranked
• In which contexts it is preferred
Recommendation scores should be interpreted as indicators of AI influence rather than customer behaviour.
11. Confidence and Consistency
AIRepScore includes two meta-indicators.AI Confidence
Confidence measures how certain AI systems appear to be about their knowledge of a brand.
High confidence does not guarantee correctness.
It indicates stronger internal certainty.
Consistency
Consistency measures the degree of agreement between different AI models.
High consistency indicates that multiple models arrive at similar conclusions.
Low consistency indicates divergent perceptions.
Consistency should be interpreted as a reliability indicator rather than a performance indicator.
12. Benchmarking Guidance
AI Positioning results should be interpreted as comparative AI perception.They do not establish factual superiority or inferiority between organisations.
Benchmarking results answer:
“How does AI position these organisations relative to one another?”
They do not answer:
“Which organisation is objectively better?”
13. Appropriate Use
AIRepScore is intended for:• Brand strategy
• Reputation management
• Competitive intelligence
• Marketing analysis
• AI visibility monitoring
• Strategic decision support
AIRepScore should not be used as the sole basis for:
• Investment decisions
• Legal decisions
• Employment decisions
• Procurement decisions
• Credit decisions
Additional sources should always be considered.
14. Methodology Evolution
The AI landscape evolves rapidly.AIRepScore is therefore maintained as a living framework.
Methodologies, model weights, supported AI systems and analytical techniques may evolve over time.
Where material methodological changes occur, AIRepScore will maintain versioning and change documentation to preserve transparency and comparability.
15. Final Disclaimer
AIRepScore measures how AI systems perceive reality.It does not determine reality.
The platform provides insight into how artificial intelligence systems currently understand, compare and recommend brands.
Users are responsible for interpreting results within the broader context of their market, organisation and business objectives.
AIRepScore™ Methodology Disclaimer & Interpretation Guide v1.0