Methodology
How the Citation Score Is Measured
CitationOS uses an AI agent-orchestrated platform to analyze how AI systems represent, cite, and recommend a firm across the major generative platforms. This page explains what we measure, how we measure it, and how scores are calculated.
Score range
0–100
A composite of three components: Citation Presence, Top-3 Rate and Narrative Depth.
Platforms measured
4
ChatGPT, Gemini, Claude and Perplexity.
Strong score threshold
70+
Firms scoring above 70 demonstrate consistent AI representation and active recommendation likelihood.
Delivery
7 days
Confidential report delivered within 7 business days of audit request submission.
§ 1
What is the Citation Score?
The Citation Score is a proprietary 0–100 metric that quantifies how effectively AI systems represent, cite, and recommend a professional service firm in high-intent decision queries. It is not a measure of website performance, search ranking, or content quality. It measures one specific thing: how AI systems interpret and act on information about a firm when a prospective client asks for a recommendation.
A firm with a Citation Score above 70 is being represented with sufficient clarity, consistency, and authority that AI systems include it in recommendations. A firm scoring below 40 has structural gaps — in entity clarity, content signals, or authority footprint — that prevent AI systems from recommending it even when the firm is technically within their knowledge base.
The Citation Score distinguishes between visibility (appearing in an AI system's knowledge base) and selection (being named in a response to a high-intent query). A firm can be visible but never selected. The Citation Score measures selection directly, and the audit identifies what causes the gap.
§ 2
How does the CitationOS platform work?
CitationOS uses an AI agent-orchestrated platform that automates the full audit pipeline from discovery to simulation. The platform does not rely on manual prompting or generic query templates. It builds a contextually accurate picture of each firm before any AI system is queried.
Step 01
Questions clients ask
High-intent prompts generated for the firm’s practice areas and market.
Step 02
AI answers across platforms
Each prompt is run on ChatGPT, Gemini, Claude and Perplexity, and every response is captured.
Step 03
Who gets named, and how
Whether the firm is mentioned, how it is described, and where it is positioned against competitors.
Step 04
Citation Score
Citation Presence, Top-3 Rate and Narrative Depth, combined into one 0–100 score.
The audit pipeline, step by step
Website crawl and service extraction
The platform crawls the firm's website and extracts its practice areas, service descriptions, market positioning, geographic focus, and key differentiators. This creates a firm-specific knowledge model that drives all subsequent analysis.
Contextual prompt generation
Using the extracted service model, the AI agent generates a set of high-intent prompts that simulate the queries a real prospective client would use when evaluating firms in the same practice area and market. These are not generic templates — they are generated specifically for each firm's competitive context.
Cross-platform simulation
The generated prompts are simulated across ChatGPT, Gemini, Claude, and Perplexity. Each platform's response is captured and analyzed for whether the firm is mentioned, how it is described, whether it is recommended, and how it is positioned relative to competitors.
Component scoring and gap identification
Responses are scored on three components: Citation Presence, Top-3 Rate, and Narrative Depth. Where the firm underperforms, the platform identifies the specific structural signals causing the gap — whether entity inconsistency, authority signal absence, content structure, or competitive displacement.
Confidential report delivery
A structured report is delivered within 7 business days. It includes the Citation Score, component-level breakdowns, platform-by-platform findings, competitive positioning analysis, and a prioritized list of corrective actions.
§ 3
What does the Citation Score measure?
The Citation Score is a composite of three components. Each component addresses a distinct layer of how AI systems name and describe a firm, and each is reported separately.
Citation Presence
How often AI systems name the firm across the full question set and all four platforms. Presence is the baseline — a necessary but insufficient condition for selection.
Example signalNamed in 12 of 40 answers.
Top-3 Rate
How often the firm is named among the first three firms in an answer. This is the component most directly connected to client acquisition.
Example signalAmong the first three names in 5 of those 12 answers.
Narrative Depth
How substantively AI systems describe the firm when they name it — from a bare name or a generic descriptor to specific practice areas, reputation, and results.
Example signalNamed, but described only as “an experienced local firm.”
Beyond the score: what the audit also examines
Representation accuracy
When the firm appears, is it described accurately? Does the AI system correctly identify the firm's practice areas, geographic focus, and positioning? Inaccurate representation actively harms selection likelihood.
Entity consistency
How consistently is the firm identified across AI platforms and across different query types? Entity inconsistency — where the same firm is described differently across platforms — is one of the most common causes of selection failure.
§ 4
Which AI platforms are included in the audit?
The Citation Score covers four of the most widely used AI assistants: ChatGPT, Gemini, Claude and Perplexity.
ChatGPT
OpenAI
Gemini
Google
Claude
Anthropic
Perplexity
Perplexity AI
Each platform uses different training data, retrieval mechanisms, and ranking logic. A firm that performs well on one platform may be absent or misrepresented on another. The audit provides a platform-by-platform breakdown so gap analysis is specific rather than averaged.
§ 5
How should the Citation Score be interpreted?
The score is designed to be actionable. Each band corresponds to a specific diagnostic profile and a set of corrective priorities.
0–29Critical
The firm is largely absent from AI recommendation pathways. Fundamental entity and content infrastructure needs to be established before platform-level optimization is meaningful.
30–49Weak
The firm appears in some contexts but is rarely recommended. Representation accuracy issues and authority signal gaps are likely present across multiple platforms.
50–69Moderate
The firm is visible but inconsistently selected. Structural gaps in entity clarity or authority signals are preventing reliable recommendation.
70–100Strong
The firm is consistently represented and actively recommended across major AI platforms. Competitive monitoring is the primary ongoing task.
§ 6
Does CitationOS manipulate or influence AI systems?
No.
CitationOS is an intelligence and diagnostic platform, not an optimization or manipulation tool. The platform observes and measures how AI systems currently represent a firm — it does not attempt to alter AI outputs, inject content into AI training pipelines, or exploit platform vulnerabilities.
The corrective actions identified in the audit report are structural improvements to a firm's digital presence — schema markup, content organization, entity clarity, authority signals — that allow AI systems to more accurately interpret what the firm does and who it serves. These are the same signals that allow AI systems to make better recommendations for prospective clients.
CitationOS does not guarantee specific AI outputs. AI systems are probabilistic and their responses vary. The audit provides a structured diagnostic of current representation and a roadmap for improving the signals that AI systems use to form recommendations.
Free snapshot
Get your free AI visibility snapshot.
We ask an AI assistant four questions a prospective client in your market might ask, and show you whether your firm is named and who is named instead.
- Four client-style questions, asked for your market
- Whether your firm is named, and which firms are named instead
- Your website’s GEO and SEO readiness
Free. Delivered as a PDF to your inbox within minutes. Privacy policy
Want the full picture? Request a confidential audit.
Thanks — your snapshot is on its way.
It usually arrives within a few minutes; if you don’t see it, check your spam or promotions folder. Websites that need a manual check can take a little longer.
Full audit
Prefer a full, confidential audit?
- How your firm appears in high-intent queries for your practice and city
- Which competitors are named instead, and why
- Your Citation Score and the specific gaps behind it
No obligation. Privacy policy
Request received.
We’ve started preparing your audit. It will arrive privately within 7 business days.