The 46 Signals That Determine Whether AI Recommends Your Law Firm

# The 46 Signals That Determine Whether AI Recommends Your Law Firm

When ChatGPT decides whether to recommend your law firm, it isn't checking your Google ranking. It's pulling from a retrieval system that evaluates 46 distinct signals across 5 layers — Trust, Entity, Retrieval, Citation, and Measurement. Most Maryland law firms have optimized for none of them.

That's not an opinion. We audited 46+ Towson-area law firms using the MAKIF-46 framework and found an average AI visibility score of 7-15 out of 100. Zero firms had a monitoring baseline. Not one was being cited by ChatGPT, Perplexity, or Google AI Overviews for the buyer-intent queries their prospective clients actually use.

This post explains what all 46 signals are, why they're weighted the way they are, and what each layer means for a Maryland law firm trying to become the AI-recommended attorney in their practice area.

Layer 1 — Trust (10 Signals, 20% Weight)

Trust is the baseline. Before any AI system retrieves or cites your content, it evaluates whether your source is credible enough to surface at all. Google's own E-E-A-T framework — Experience, Expertise, Authoritativeness, Trustworthiness — explicitly identifies trust as the foundational dimension that all other quality signals depend on.

For law firms, Trust layer signals include:

Research published at KDD '24 by Princeton and Columbia found that adding statistics and sourced citations to content produces approximately 32% improvement in AI visibility. <cite><a href="https://arxiv.org/abs/2406.13692" rel="external">Aggarwal et al., KDD '24, arxiv.org/abs/2406.13692</a></cite> For Law & Government queries specifically, Statistics Addition was the single highest-performing optimization method tested. Trust layer signals are the foundation that makes those citations credible.

Layer 2 — Entity (7 Signals, 20% Weight)

An entity is not a keyword. It is a node in a knowledge graph — a specific, disambiguated thing with defined properties and relationships. Google's Knowledge Graph contains millions of entities. Your law firm is either in it, with a clear identity, or it is noise.

Entity layer signals determine whether AI systems know *who you are* and can confidently attribute content to your specific firm.

In our Towson audits, the Entity layer averaged 5-15 out of 100. Entity confusion — AI systems uncertain whether the business is a real, specific, locatable firm — is the most common reason a law firm is retrieved but then excluded from the final AI response.

Layer 3 — Retrieval (8 Signals, 30% Weight)

Retrieval gets the highest weight (30%) because it is the prerequisite for everything else. Before AI can cite you, it has to find you. Modern AI search systems use Retrieval-Augmented Generation (RAG) architectures that pull content from external sources at query time. Your content has to be structured in a way that matches how these systems select documents.

Research published at KDD '24 found that adding relevant statistics to content produces approximately a 32% AI visibility improvement. <cite><a href="https://arxiv.org/abs/2406.13692" rel="external">Aggarwal et al., KDD '24, arxiv.org/abs/2406.13692</a></cite> But that only works if the content is retrieved in the first place.

Consider the DUI practice area page that answers "What is Probation Before Judgment Maryland DUI?" with a direct answer in the first paragraph, a citation to Maryland Courts and Judicial Proceedings Article, FAQ schema with 8 specific Q&As, and 1,800+ words of Maryland-specific content. That page has strong R3, R6, R7, and R8 signals. It gets retrieved. The 400-word generic "DUI Defense" page gets skipped.

Layer 4 — Citation (6 Signals, 20% Weight)

Getting retrieved doesn't guarantee getting cited. The Citation layer is where content quality has its most direct, research-validated impact on AI visibility.

The Citation layer is where most Towson law firms score 0-10. There is nothing on their sites worth quoting. No statistics. No statute section numbers. No specific deadlines. No declarative legal statements with named sources.

Layer 5 — Measurement (4 Signals, 10% Weight)

Only 16% of businesses currently track their AI search performance in any systematic way. Among the 46+ Towson law firms we audited, the number tracking AI citations was zero. Zero out of 46.

Without Measurement, you cannot know whether your GEO investments are working. You cannot see your competitors gaining ground. You cannot identify which specific queries you've won or lost. Measurement is the lightest layer by weight (10%) but the one that determines whether all other investments compound over time.

The Magnificent 7 — Your Practice Area's 7 Buyer-Intent Queries

Every practice area has seven high-intent queries that represent the questions real clients ask at the moment of need. We call these the Magnificent 7. They are the specific prompts AI systems are asked most frequently for your practice area — and the ones your firm should be cited for.

For Personal Injury, the Magnificent 7 are:

1. "What to do after a car accident in Maryland"

2. "Maryland contributory negligence personal injury"

3. "How much is my Maryland car accident worth"

4. "How do contingency fees work Maryland"

5. "Third-party claim workers comp Maryland"

6. "Statute of limitations personal injury Maryland"

7. "Personal injury lawyer Towson MD"

When we ran all seven of these queries across ChatGPT, Perplexity, and Google AI Overviews in our Towson audit — zero local personal injury firms were cited for any of them. The cited sources were national legal information sites: Nolo, FindLaw's editorial content, Justia's legal guides. Not a single Towson PI firm.

Every practice area has its own Magnificent 7. Criminal defense. Family law. Estate planning. Employment law. The pattern is the same: no Towson firm is winning any of them.

Why Retrieval Gets 30% Weight

The 30% weight assigned to the Retrieval layer reflects the research finding that retrieval failure is the most common and most silent failure mode in AI visibility. A firm can have strong Trust signals (real attorneys, real credentials) and strong Citation signals (statute references, statistics) — but if the content is never retrieved, none of it matters.

FAQ schema significantly increases citation probability by structuring content as direct question-answer pairs that AI retrieval systems are optimized to extract — a core principle of the MAKIF-46 Retrieval layer. The KDD '24 research found that fluency and readability improvements produce approximately a 31% AI visibility boost. <cite><a href="https://arxiv.org/abs/2406.13692" rel="external">Aggarwal et al., KDD '24, arxiv.org/abs/2406.13692</a></cite> Content that is both structured (FAQ schema) and well-written (high fluency) sits at the intersection of the two highest-performing retrieval interventions.

Frequently Asked Questions

*What is the MAKIF-46 framework?*

The MAKIF-46 is a 46-signal audit framework that evaluates law firm AI visibility across five weighted layers: Trust (10 signals, 20%), Entity (7 signals, 20%), Retrieval (8 signals, 30%), Citation (6 signals, 20%), and Measurement (4 signals, 10%). Scores range from 0-20 (INVISIBLE) to 81-100 (AI-DOMINANT). The average Towson law firm scores 7-15.

*Why does Retrieval get the highest weight at 30%?*

Because retrieval failure is the most common and most consequential failure mode. Strong Trust and Citation signals are irrelevant if AI systems never retrieve your content. The 30% weight reflects both the frequency of retrieval failure in our audits and the magnitude of its impact on downstream citation rates.

*What is FAQ schema and why does it matter?*

FAQ schema is structured markup that labels Q&A content so AI systems can identify and retrieve it specifically as question-answer pairs. FAQ schema significantly increases citation probability by structuring content as direct question-answer pairs that AI retrieval systems are optimized to extract. It's the highest-leverage single technical change most law firms can make.

*How is the MAKIF-46 different from a standard SEO audit?*

A standard SEO audit evaluates ranking signals for Google's PageRank algorithm: backlinks, keyword density, page speed, meta tags. The MAKIF-46 evaluates signals that determine AI citation: entity clarity, retrieval structure, citation triggers, content freshness, and measurement infrastructure. Only 12% of AI-cited URLs overlap with Google's top 10. <cite><a href="https://arxiv.org/abs/2406.13692" rel="external">KDD '24 Research, 2026</a></cite> The two systems require different optimization strategies.

*What are the Magnificent 7?*

The Magnificent 7 are the seven highest buyer-intent queries for a given practice area — the questions prospective clients ask AI systems at the moment they need legal help. Every practice area has its own Magnificent 7. When we ran the Personal Injury Magnificent 7 across ChatGPT, Perplexity, and Google AI Overviews in Towson, zero local PI firms were cited for any of them.

*How long does it take to improve a law firm's MAKIF-46 score?*

Entity fixes (schema, Business Profile, NAP consistency) typically show results in 4-6 weeks. Retrieval fixes (FAQ schema, content structure, llms.txt) show results in 6-8 weeks. Citation improvements (statistics integration, statute references, content depth) take 60-90 days to propagate. A full 90-day GEO program can realistically move a firm from INVISIBLE (0-20) to CITATION-ELIGIBLE (61-80).

*Does having a high Google ranking help with AI visibility?*

Minimally. Only 12% of AI-cited URLs overlap with Google's top 10. <cite><a href="https://arxiv.org/abs/2406.13692" rel="external">KDD '24 Research, 2026</a></cite> Strong technical SEO (HTTPS, clean sitemaps, mobile responsiveness) supports AI crawlability, but keyword rankings and backlink profiles have minimal direct impact on AI citation rates. The optimization strategies are distinct.

*What should a Maryland law firm do first to improve AI visibility?*

In priority order: (1) Fix schema to 'LegalService' with attorney sub-type, (2) implement FAQ schema with Maryland-specific legal Q&As citing statutes by section number, (3) add named attorney attribution to every content page, (4) audit and fix NAP consistency across all directories, (5) establish weekly AI query monitoring baseline. These five interventions address the most consistently failing signals in the Towson market.

Want your firm's MAKIF-46 score? [Book the Audit](/audit) and see where you stand against the rest of the Towson legal market.

Sources: KDD '24 GEO Research (Aggarwal et al., arxiv.org/abs/2406.13692) · SE Ranking AI Overviews Research 2024 · WebFX / Semrush AI Referral Traffic Research · McKinsey AI Search Report 2025 · MAKIF Audit Data, Towson MD, 2025-2026