Larry H. Miller Honda Murray

The short answer

LHM Honda Murray, one of the Larry H. Miller dealerships within the Asbury Automotive Group network, had a 9.4% AI visibility score. That means for over 90% of the AI-generated answers about Honda vehicles and Honda dealerships in the Salt Lake City market, the dealership was not being cited, referenced, or even mentioned. It was functionally invisible in the fastest-growing search channel in automotive retail.

Within one month of implementing a structured SEO, AEO, and GEO strategy, the visibility score exceeded 40%. The dealership became the #1 cited source on ChatGPT, Perplexity, Google AI Overviews, and Claude for a wide range of Honda-related queries in the Salt Lake City DMA.

That includes outranking Honda's own corporate website on the top five most-searched Honda questions nationally. Honda corporate has a massive domain authority advantage, dedicated content teams, and decades of brand equity. Outranking them required the content to be structurally superior in terms of entity clarity, answer completeness, and citation accessibility.

This project draws directly from the SEO, AEO, and GEO framework I architect and implement across the broader Asbury Automotive Group dealership network.

Problem

The shift from search to AI-powered discovery

The automotive industry is experiencing a fundamental change in how consumers research and purchase vehicles. For two decades, the buyer journey started with a Google search: "best Honda SUV 2025," "Honda dealers near me," "Honda CR-V vs Toyota RAV4." Dealers invested heavily in traditional SEO and paid search to capture these queries, and the ones who did it well generated a steady stream of leads.

That model is not disappearing, but it is being supplemented and increasingly replaced by AI-powered research. A growing share of car buyers now start their research by asking an AI tool a question: "What is the best Honda dealer in Salt Lake City?" "Is the 2025 Honda CR-V worth buying?" "Which Honda models have the best resale value?"

What a 9.4% visibility score means

I use tools like Profound to measure AI visibility scores. The score represents the percentage of relevant AI-generated responses in which a specific entity is cited or mentioned. A 9.4% score means that out of all the AI-generated answers to Honda-related queries in the Salt Lake City DMA, the dealership appeared in fewer than one in ten.

To understand why this matters, consider what was getting cited instead:

• Honda corporate (honda.com) was the most frequently cited source for general Honda model information, specifications, and pricing.
• Third-party aggregators like Edmunds, Cars.com, KBB, and CarGurus were being cited for comparison content, reviews, and pricing guidance.
• Competing dealerships in the Salt Lake City market with more structured content and better entity markup were appearing in local queries.
• LHM Honda Murray was being passed over despite having competitive pricing, strong inventory, a highly-rated service department, and a long track record in the market.

The dealership had the substance to be the authoritative source. What it lacked was the technical optimization that would allow AI models to recognize, extract, and cite its content.

Why traditional SEO alone was not enough

The dealership had invested in traditional SEO for years. It ranked reasonably well in conventional Google results for core transactional queries. But traditional SEO is optimized for a fundamentally different information retrieval model.

Traditional SEO is about page-level optimization: getting a specific URL to rank for a specific keyword. You optimize title tags, meta descriptions, header tags, and page content around target keywords. The unit of optimization is the page, and the goal is to get that page onto the first page of search results.

AI-powered search works differently. AI models do not return pages. They return answers. The model reads hundreds or thousands of pages, extracts relevant information, synthesizes it, and presents a coherent response. The sources that get cited are the ones the model identifies as authoritative, complete, and easy to extract structured answers from.

Approach

The strategy was built on three pillars that work together. Each pillar addresses a different dimension of how AI models discover, evaluate, and cite content.

Before optimizing for AI citation, the traditional SEO foundation needed to be solid. AI models consume web content through crawling and indexing pipelines that share many characteristics with traditional search engine crawlers. If your content is not crawlable, indexable, and structured in a way that search systems can parse, it will not be available for AI models to evaluate either.

Technical Audit and Remediation

I audited the site's technical SEO health: crawlability, indexation coverage, site speed, mobile usability, URL structure, internal linking architecture, and structured data implementation. The site had several technical issues that were limiting its crawl efficiency and indexation depth. Pages that were not being indexed could not be evaluated by AI models. Pages that loaded slowly or had rendering issues were less likely to be fully parsed during crawl. These foundational issues needed to be resolved before any content optimization would have full impact.

Content Architecture Restructure

The site's content was reorganized around intent clusters rather than individual keywords. Instead of having a single model page for the Honda CR-V that tried to cover everything, the content was broken into distinct pages targeting distinct intents: pricing and financing, features and specifications, comparison against competitors, owner reviews and satisfaction, and local availability. Each page was designed to be the definitive resource for its specific intent.

Internal Linking and Topic Authority

Internal links were restructured to create clear topical hierarchies. The Honda CR-V pricing page links to the CR-V features page, which links to the CR-V comparison page, which links to the CR-V reviews page. This creates a cluster of interconnected pages that signals to both search engines and AI models that the site has comprehensive, authoritative coverage of the topic. A site with five well-linked, intent-specific pages on a model demonstrates more topical authority than a site with one page that tries to cover everything.

AEO is the discipline of optimizing content specifically for citation by AI answer engines. It is the most technically distinct of the three pillars because it requires understanding how AI models select sources for citation, which is fundamentally different from how traditional search engines rank pages.

Entity-First Content Architecture

The most important concept in AEO is entity recognition. AI models do not think in terms of pages and keywords. They think in terms of entities and relationships. An "entity" is a distinct, identifiable thing: a business, a vehicle model, a geographic market, a service offering. AI models build internal representations of entities and their attributes, and they cite sources that provide clear, structured information about those entities.

I restructured the dealership's content so that AI models could identify LHM Honda Murray as a distinct entity with clearly defined attributes and relationships:

• What the entity is: A Honda dealership, part of the Larry H. Miller brand, within the Asbury Automotive Group network, located in Murray, Utah, serving the Salt Lake City DMA.
• What the entity offers: New Honda vehicle sales, certified pre-owned vehicles, service and maintenance, parts, financing, specific model inventory.
• How the entity relates to other entities: Relationship to Honda as a manufacturer, relationship to specific vehicle models, relationship to the geographic market, relationship to competing dealerships.
• What makes the entity authoritative: Customer satisfaction ratings, service department certifications, inventory depth, pricing competitiveness, years in operation, community involvement.

Every page on the site was restructured to reinforce this entity definition. The goal was that an AI model reading any page on the site could build a complete, accurate picture of what LHM Honda Murray is, what it offers, and why it is a credible source of information.

Answer-Formatted Content

AI models select sources for citation based on how easily they can extract a clean, complete answer from the content. A page full of marketing copy and promotional language is harder for a model to extract a factual answer from than a page that directly addresses a specific question with structured, factual content.

I identified the specific questions that AI models were being asked about Honda vehicles in the Salt Lake City market and created content designed to be the definitive answer to each one:

For each of these queries and dozens more, the content was structured so that the answer could be extracted in one or two clear passages. AI models prefer sources where they do not have to synthesize across multiple sections of a page to construct a response. The easier you make extraction, the more likely you are to be cited.

Structured Data and Schema Graph

I implemented comprehensive structured data that went significantly beyond standard automotive dealer markup. The schema graph connected the dealership entity to its inventory, service department, geographic market, business attributes, and content assets in a machine-readable format.

The schema implementation included:

GEO ensures that content is recognized as locally authoritative rather than just generically relevant. This distinction matters enormously for automotive retail because car buying is an inherently local activity. A customer in Salt Lake City asking "What is the best Honda dealer?" is not looking for a generic national answer. They want to know which dealer in their market is the best option.

Local Entity Signals

I built geographic authority signals into every layer of the optimization:

• Market-specific content that references the Salt Lake City DMA, surrounding communities, and local landmarks. Content that mentions "serving Murray, Salt Lake City, West Jordan, Sandy, and the greater Wasatch Front" establishes geographic scope in a way that AI models can parse.
• Location-aware structured data that explicitly defines the dealership's service area, geographic coordinates, and market context. AI models use this data to match the dealership to queries with geographic intent.
• Local citation consistency across all external directories, profiles, and listings. Consistent NAP data across Google Business Profile, Bing Places, Yelp, DealerRater, and automotive-specific directories reinforces the dealership's local entity identity.

Local Link and Citation Signals

Local authority is reinforced by the quality and relevance of external signals pointing to the dealership. I identified opportunities for local citations, community involvement mentions, local business directory listings, and market-specific content partnerships. Each of these signals tells AI models that the dealership is an established, recognized entity within its specific geographic market.

Query-Specific Local Content

For every high-intent local query pattern ("Honda dealer near me," "best Honda dealer in Salt Lake City," "Honda service Salt Lake City"), I created content specifically designed to be the definitive local answer. This content combines the entity definition from AEO, the structured data from the schema implementation, and the geographic signals from GEO into a comprehensive, citation-ready response to each query.

Result

The speed of the results

The results were not gradual. Traditional SEO improvements typically compound over months as search engines re-crawl, re-index, and re-evaluate content. AI visibility moved faster for a specific reason: AI models are constantly re-evaluating their source selections based on content quality, structured data availability, and entity clarity. When a site goes from having no structured data and no entity-optimized content to having comprehensive schema markup and citation-formatted content, the AI model's evaluation of that site's authority can shift in a single re-indexing cycle.

The visibility score showed measurable improvement within the first two weeks. By the end of the first month, the dealership had moved from invisible to dominant in its local market across every major AI platform.

Outranking Honda corporate

Honda corporate has a domain authority that no individual dealership can match. Honda's website has millions of backlinks, decades of content history, and a brand recognition advantage. In traditional SEO, competing with Honda corporate on broad Honda queries would be nearly impossible.

But AI citation selection does not work purely on domain authority. It works on answer quality, entity specificity, and structured data completeness. This is the core insight of the SEO + AEO + GEO framework: you do not need to be the biggest to be the most citable. You need to be the most structured, the most specific, and the most complete for the queries that matter to your business.

In the client's words

"In one month, Zach took the dealership from being invisible in AI search to being the #1 cited source for Honda questions in Salt Lake City. That included outranking Honda's own corporate website on the top five most-searched Honda questions nationally. What he built here has become the model we're now applying across the entire dealership network."

— General Manager, Larry H. Miller Honda Murray

What I'd do differently

Implement AI citation share tracking from day one as a leading indicator alongside the visibility score. The visibility score is the headline metric: it tells you the percentage of relevant AI-generated answers where you are cited. But citation share data from tools like Profound gives earlier, more granular signal on which specific content changes are driving the most impact. You can see which pages are being cited, for which queries, and how frequently, before those changes fully roll up into the aggregate visibility score.

Having that granular data from the start would have allowed me to double down on the highest-performing optimizations even faster during the critical first two weeks. For example, the FAQ schema implementation drove an outsized share of early citation gains, but I did not have the query-level data to confirm that until week three. If I had been tracking citation share from day one, I could have expanded the FAQ content earlier and accelerated the overall visibility trajectory.