A GEO Score is a single number from 0 to 100, but that number is built from three distinct components: frequency (40% weight), prominence (35% weight), and breadth (25% weight). Each component measures a different dimension of your law firm’s AI visibility across ChatGPT, Claude, Gemini, and Perplexity. Understanding what each component measures, how it is calculated, and what drives it up or down is essential for developing an effective optimization strategy. This article is the technical deep dive,the companion piece to our overview of what a GEO Score is and why it matters.

The Three Dimensions of AI Visibility

When Pinniq Legal runs a GEO scan for your law firm, it fires dozens of realistic, client-style queries at all four major AI platforms. Each query is the kind of question a real potential client would ask: “Who is the best personal injury lawyer in Nassau County?” “I need a car accident attorney in Mineola, who do you recommend?” “Which law firms handle trucking accident cases in Long Island?”

For each query on each platform, Pinniq Legal records whether your firm was mentioned, where it appeared in the response (first, second, third, etc.), and which platform generated the response. These raw data points are then aggregated into three component scores:

  • Frequency measures how often your firm is mentioned across all queries and platforms.
  • Prominence measures where your firm appears when it is mentioned,first recommendation, third, or fifth.
  • Breadth measures how many of the four AI platforms recommend your firm.

Each component captures something the others do not. A firm can be mentioned frequently but always as the last recommendation (high frequency, low prominence). A firm can be the first recommendation on ChatGPT but invisible on Gemini, Claude, and Perplexity (high prominence, low breadth). Understanding your score at the component level reveals exactly where your AI visibility strengths and weaknesses lie.

Frequency: How Often You Are Mentioned (40% Weight)

Frequency is the foundation of your GEO Score and carries the largest weight at 40%. It answers the most fundamental question: when potential clients ask AI assistants for attorney recommendations in your practice area and geography, how often does your firm appear in the response?

How frequency is calculated

Frequency is expressed as a percentage. If Pinniq Legal fires 40 total queries across all four platforms and your firm appears in 12 of those responses, your raw frequency is 30% (12 / 40). This percentage is then scaled to a 0-100 component score.

The calculation accounts for all query types in your scan: broad practice area queries (“best personal injury lawyer in [county]”), sub-type queries (“car accident attorney in [county]”, “medical malpractice lawyer in [county]”), and town-level queries (“personal injury attorney in [town]”) for plans that include town-level data.

What a good frequency score looks like

For most law firms running their first scan, frequency is the weakest component. The median frequency for new Pinniq Legal clients is approximately 15%,meaning the firm appears in only about 1 in 7 AI responses. Here is how to interpret your frequency score:

  • 0-10%: Your firm is essentially invisible to AI. Most queries return no mention of your firm on any platform.
  • 11-30%: Sporadic mentions. You appear on some platforms for some queries but lack consistency.
  • 31-50%: Moderate frequency. AI mentions you regularly but not in the majority of relevant queries.
  • 51-75%: Strong frequency. You are a consistent recommendation across most query types.
  • 76-100%: Dominant frequency. AI mentions your firm in nearly every relevant query. Very few firms achieve this.

How to improve frequency

Frequency improves when AI models have more reasons to associate your firm with the queries being asked. The most impactful strategies are:

  • Directory consistency: Ensure your firm appears on every major legal directory (Avvo, FindLaw, Justia, Super Lawyers, Martindale-Hubbell, state bar) with consistent name, address, and practice area listings. Each directory is an additional data point that AI models can draw from.
  • Review volume: AI models use review count and recency as a relevance signal. Firms with more recent reviews mentioning their practice area and location are mentioned more frequently.
  • Practice area content: Dedicated practice area pages on your website with clear, entity-rich content (firm name + practice area + location) help AI models match your firm to relevant queries.
  • Sub-type coverage: Many firms are mentioned for their primary practice area but invisible for sub-types. A personal injury firm might appear for “personal injury lawyer” but not for “truck accident attorney” or “wrongful death lawyer.” Creating dedicated content for each sub-type closes these gaps. Read our guide to content changes that improve GEO for specific tactics.
Key Takeaway: Frequency is the highest-weighted component at 40% and the area where most firms have the most room for improvement. The primary lever is building consistent citations across authoritative sources so AI models have abundant evidence to draw from when generating recommendations.

Prominence: Where You Appear in the Response (35% Weight)

Prominence measures the quality of your mentions, not just the quantity. When an AI assistant lists 3 to 5 law firms in response to a query, being the first firm named is significantly more valuable than being the fifth. Users pay the most attention to the first recommendation and are most likely to contact that firm.

How prominence is calculated

For each query where your firm is mentioned, Pinniq Legal assigns a position-based quality score:

  • First mention: 100 points. Your firm is the first attorney recommendation in the response.
  • Position 2-3: 60 points. Your firm appears early in the response, in the top tier of recommendations.
  • Position 4-5: 30 points. Your firm is mentioned but in the lower half of the typical recommendation list.
  • Position 6+: 10 points. Your firm is mentioned but buried deep in the response, where fewer users will see it.

Your prominence score is the average of these position scores across all queries where your firm was mentioned. If you were first in 5 queries (500 points), third in 3 queries (180 points), and fifth in 2 queries (60 points), your average prominence is 74 out of 100 (740 / 10).

Note that prominence is only calculated for queries where your firm appears. If you are not mentioned at all, that query does not affect your prominence score,it affects your frequency score instead. This distinction is important: prominence measures how well you perform when you do appear, while frequency measures how often you appear at all.

What a good prominence score looks like

  • 0-25: When you are mentioned, you are typically listed as the 5th or later recommendation. Users rarely make it this far in an AI response.
  • 26-50: You appear in the middle of recommendation lists. Some users will see you, but you are not the firm AI leads with.
  • 51-75: You are consistently in the top 3 recommendations. This is strong performance that directly translates to user engagement.
  • 76-100: You are frequently the first or second firm recommended. This is the strongest possible prominence position.

How to improve prominence

Prominence is harder to influence than frequency because it depends on AI models perceiving your firm as the best answer, not just a valid one. The strategies that improve prominence are:

  • Published case results and verdicts: AI models that cite specific dollar amounts and case types tend to list the firm with the most impressive results first. Publishing detailed verdicts and settlements on your website gives AI models concrete reasons to feature you prominently.
  • Awards and recognitions: Super Lawyers, Best Lawyers, National Trial Lawyers,these named recognitions serve as ranking signals for AI models deciding the order of recommendations.
  • Media coverage: Firms that appear in news articles, legal publications, and bar association features tend to rank higher in prominence because these sources carry high authority weight in AI models.
  • Specialization depth: AI models tend to rank specialized firms higher than generalists for specific queries. A firm that is clearly the authority on truck accident cases in a specific county will often outrank a larger general practice firm in prominence for truck accident queries.

Breadth: How Many Platforms Recommend You (25% Weight)

Breadth measures the diversity of your AI presence across the four major platforms: ChatGPT, Claude, Gemini, and Perplexity. It carries the smallest weight at 25%, but it serves a critical strategic function: it measures your resilience against platform-specific changes.

How breadth is calculated

Breadth is a simple metric. Pinniq Legal tracks which of the four platforms mention your firm in at least one response. If your firm appears on all four platforms, your breadth score is 100. If you appear on three platforms, your breadth is 75. Two platforms gives you 50, one platform gives you 25, and zero platforms gives you 0.

This is a binary measure per platform,the platform either recommends you or it does not. The number of times it recommends you on each platform does not affect breadth (that is captured in frequency). Breadth strictly measures cross-platform coverage.

What a good breadth score looks like

  • 0-25: You appear on zero or one platform. Your AI visibility is extremely fragile and concentrated.
  • 50: You appear on two of four platforms. You have some coverage but significant gaps.
  • 75: You appear on three of four platforms. This is the most common score for firms with moderate AI visibility. The missing platform is usually Claude, which lacks web search.
  • 100: You appear on all four platforms. Maximum breadth provides the most stable and resilient AI presence.

How to improve breadth

The most common breadth gap is Claude. Because Claude does not use web search and relies entirely on training data, firms that are visible on ChatGPT, Gemini, and Perplexity (all of which use web search) often score 0 on Claude. Closing this gap requires a different approach than the other three platforms:

  • For Claude: Build broad, authoritative online presence across sources that are likely included in AI training datasets. Legal directories, bar association pages, published legal articles, and media coverage all contribute to training data. Claude visibility is a long-term investment that pays off when new model versions are released.
  • For ChatGPT: Ensure your website and directory listings are crawlable and up to date. ChatGPT’s web search tool needs to find current, relevant content about your firm.
  • For Gemini: Optimize your Google Business Profile, Google reviews, and local SEO signals. Gemini’s Google Search grounding means your Google ecosystem presence directly influences recommendations.
  • For Perplexity: Maximize the number of authoritative sources that mention your firm. Perplexity searches widely and cites its sources, so appearing on many reputable sites increases your chances of being recommended and cited.
Key Takeaway: Breadth protects you against platform-specific changes. If ChatGPT updates its model and your mentions drop, having strong visibility on Gemini, Perplexity, and Claude means you still have three channels generating referrals. The most common gap is Claude, which requires long-term authority building rather than quick content updates.

Per-Platform Scoring: Diagnosing Specific Weaknesses

In addition to the three-component composite GEO Score, Pinniq Legal provides per-platform mention rates for each of the four AI platforms. These are straightforward percentages: the number of queries where your firm was mentioned on a specific platform, divided by the total queries sent to that platform.

Per-platform scores are essential for diagnosis. Consider two firms that both have a frequency score of 40%:

  • Firm A: ChatGPT 60%, Gemini 55%, Perplexity 45%, Claude 0%. This firm has strong web-search visibility but no training-data presence. Their optimization priority should be building long-term authority signals for Claude.
  • Firm B: ChatGPT 25%, Gemini 65%, Perplexity 30%, Claude 40%. This firm has unusually strong Claude and Gemini presence but weak ChatGPT and Perplexity visibility. Their priority should be fresh, crawlable content and source diversity for the web-search platforms.

Without per-platform data, both firms would receive the same advice. With it, each gets a targeted strategy for their specific weaknesses. This is why Pinniq Legal reports per-platform scores alongside the composite GEO Score,the composite tells you where you stand overall, and the platform breakdowns tell you where to focus your optimization efforts.

How the Composite Score Works

The final GEO Score is a weighted composite of the three components:

GEO Score = (Frequency × 0.40) + (Prominence × 0.35) + (Breadth × 0.25)

This weighting reflects the relative importance of each dimension for law firm AI visibility:

  • Frequency at 40% carries the most weight because being mentioned at all is the prerequisite for everything else. A firm that is never mentioned has zero AI visibility regardless of how prominently it would appear if it were mentioned.
  • Prominence at 35% is the second-highest weight because position directly impacts whether users contact a firm. Being the first recommendation converts dramatically better than being the fifth.
  • Breadth at 25% carries the smallest weight because it is a stability metric. Cross-platform presence is important for resilience, but frequency and prominence on even one platform can generate meaningful referrals.

For scans that include town-level data (Intelligence and Managed plans), the overall GEO Score also incorporates a territory weighting: 60% of the score comes from county-level (PMT) queries and 40% comes from the average of town-level (TRN) query scores. This reflects the reality that county-level queries are more competitive and represent higher-value searches, while town-level queries provide granular local visibility.

Common Score Patterns and What They Mean

After scanning hundreds of law firms, certain GEO Score patterns appear repeatedly. Recognizing your pattern helps you prioritize the right optimization strategy:

Pattern 1: Low everything (GEO Score 0-10)

The firm is essentially invisible to AI. This is the most common pattern for firms that have never optimized for AI visibility. The fix starts with the fundamentals: directory consistency, structured content, and review generation.

Pattern 2: Moderate frequency, low prominence (GEO Score 15-30)

AI mentions the firm occasionally but always late in the recommendation list. The firm has basic visibility signals (directory listings, some reviews) but lacks the authority signals that make AI models prioritize it. The fix: publish case results, earn named recognitions, and build media coverage.

Pattern 3: High frequency on one platform, zero on others (GEO Score 20-35)

The firm has strong visibility on Gemini (because of good Google optimization) but is invisible on ChatGPT, Perplexity, and Claude. Or it appears on ChatGPT but nowhere else. The fix: diversify your citation sources and build cross-platform authority so your presence is not dependent on any single platform’s data pipeline.

Pattern 4: Strong broad score, weak sub-type scores (GEO Score 30-45)

The firm appears for broad practice area queries but is invisible for specific sub-types like truck accidents, medical malpractice, or wrongful death. The fix: create dedicated content and directory listings for each practice area sub-type, and generate reviews that mention specific case types.

Pattern 5: Strong county score, weak town scores (GEO Score 35-50)

The firm dominates county-level AI queries but is invisible at the town level. This is common for firms in competitive markets where county-level competition is fierce but town-level queries are wide open. The fix: add town-specific content, citations, and Google Business Profile service areas for the individual towns in your market.

Key Takeaway: Your composite GEO Score tells you how visible you are overall. The component breakdown (frequency, prominence, breadth) and per-platform scores tell you exactly why and where to focus. Most firms fall into one of five common patterns, each with a specific optimization playbook. Run your free scan to see which pattern matches your firm.