How a 2-Year-Old Roofing Company Gets More AI Mentions Than You and How to Fix It

You've been in business 18 years. You rank on Google. You have 200 reviews. And a roofing company that opened in 2023 just got recommended by ChatGPT to a homeowner in your zip code whose roof is leaking right now.

That homeowner will never see your Google ranking.

This isn't a failure of your SEO. It's a failure to understand what AI actually rewards, and it's nothing like what you've been optimizing for.

The Uncomfortable Truth: AI Doesn't Read Your Google Rankings

You've built a solid SEO house. Top 3 positions, steady organic traffic, years of domain authority. Great. But none of that matters to ChatGPT, Claude, Perplexity, or Gemini when someone asks for a recommendation.

Here's the mechanism: LLMs don't crawl Google SERPs. They train on the raw internet and then use structured data patterns, entity signals, and mention volume to determine what to cite.

Ahrefs' 2025 data found something striking: 80% of content cited by LLMs doesn't rank in Google's top 100. The two algorithms are parallel systems, not connected ones. You can dominate Google and be invisible to AI simultaneously.

Your 18 years of SEO equity transferred to AI relevance? Maybe 20 to 30%. Your meta descriptions, keyword-optimized service pages, and backlink profile? Almost worthless to an LLM. Your schema markup, review signals, and entity consistency? Everything.

Why National Chains and New Companies Beat You in AI

HomeAdvisor, Angi, and Thumbtack have 50 million+ pages indexed. When a homeowner asks "best roofer near me," an LLM is pulling from a massive volume of structured mentions across those platforms. They're not asking Google who ranks first. They're calculating: which company appears most consistently across trusted sources, with the cleanest data, the most service detail, and the strongest review signals?

New roofing companies that launched in 2022 to 2023 have an advantage you didn't plan for: they built their digital footprint after LLMs were trained. Their Google My Business profile was set up with proper schema. Their website service pages were built with semantic clarity. Their citations across directories included NAP consistency from day one. Their review text was captured in structured formats.

Your site was built for a 2010 search paradigm. You've SEO'd it forward, but not AI'd it forward.

The 2-year-old company doesn't need to be better. They just need to be shaped for AI. And that shape is radically different from Google's shape. Their Google My Business profile was set up complete and consistent from day one. Their website schema was built for AI readability. Yours was built for Google crawlers.

What AI Actually Reads When Someone Asks "Best Roofer Near Me"

An LLM isn't reading your title tags or H1s. It's reading a specific data layer:

Schema markup. This is where you separate yourself from that 2-year-old competitor. Most roofing sites have basic LocalBusiness and Service schema. You need HasOfferCatalog.

Here's the mechanism: HasOfferCatalog allows you to list every single service, material type, and repair style in a nested, machine-readable structure. Instead of AI seeing "roofing company," it sees:

Company X offers:
├─ Asphalt Shingle Roof Installation
│  ├─ Architectural shingles
│  ├─ 3-tab shingles
│  └─ Impact-resistant shingles
├─ Metal Roof Installation
├─ Flat Roof Repair
│  ├─ Torch-down repair
│  ├─ Membrane sealing
│  └─ Leak detection
├─ Storm Damage Assessment
└─ Gutter Installation & Repair
   ├─ Seamless gutters
   ├─ Guard installation
   └─ Downspout extensions

When ChatGPT is asked "I have a 20-year-old asphalt roof and it's leaking. What are my options?", it can now match your specific expertise to that query. The 2-year-old upstart with generic "Roofing" schema gets skipped. You get cited.

Service pages. AI is looking for semantic clarity on what you offer, service areas, pricing range if disclosed, and how you describe your process. Vague pages ("We provide roofing services") disappear. Clear pages ("We specialize in architectural shingle installation, torch-down flat roofing, and emergency storm repair in a 25-mile radius") get cited.

Review signals. Not just the star rating. The actual text of reviews. AI pulls language patterns, service mentions, and proof of claims from review content. A review saying "fixed a leak on my old asphalt roof, quick turnaround" is a data point. 200 reviews with that pattern is a signal.

Directory signals. Google My Business, Yelp, HomeAdvisor, contractor boards, and consistency across them. Mismatched phone numbers, different service area descriptions, or missing locations cost you.

NAP consistency. Name, address, phone. Across all platforms. A roofing company with three different phone numbers across directories is algorithmically weaker than one with one number everywhere. Mismatched listings actively suppress your AI visibility, and most contractors don't know it's happening.

Entity recognition. Your company name, service keywords, location, and how they co-occur across the web. An LLM is pattern-matching: this entity (your company) is associated with these services in this location and these other entities link to it.

Compare this to what you've optimized for: keyword density in your service pages, backlink anchor text, page speed. You've optimized for Google's crawl. You've ignored AI's read.

What Transfers From Your 18 Years of SEO and What Doesn't

Transfers:

• Domain authority. Your old site has been around. That's a signal. LLMs factor in long-term web presence.
• Review volume and ratings. 200 five-star reviews is a massive advantage over a startup with 15. AI weights this heavily.
• Google My Business foundation. If you've been maintaining it, your location data is established.

Doesn't transfer:

• Keyword-stuffed service pages. "Roofing services in Denver, roofing contractor Denver, Denver roofer, roof repair Denver": an LLM reads this as low-signal content.
• Thin city pages. 300 words about "roofing in [City Name]" with minimal unique information. AI skips these.
• Meta descriptions and title tag optimization. Completely invisible to LLMs.
• Traditional backlinking strategy. Unless those links include structured entity signals or mentions of your service offerings, an LLM doesn't care much.
• High Google rankings. The two systems are independent. Ranking #1 on Google for "roofing contractor Denver" doesn't make you #1 for LLM recommendations.

The painful part? You've invested heavily in what doesn't transfer. The good news? What does transfer is already yours. You just need to build on top of it.

The 3-Month Retrofit: 60 Days to AI Visibility (No Tools Required)

You don't need new software. You need a structural rebuild that takes most contractors 2 to 3 months. Here's the exact order:

Month 1: Schema and Service Pages

Week 1 to 2: Audit your current schema. Go to your homepage in a browser, right-click, view page source, search for "schema.org." Note what's missing. Most roofing sites, including your competitor, have no schema or incomplete LocalBusiness markup.

Here's your advantage: the 2-year-old competitor probably has basic schema. You're going deeper. You're implementing HasOfferCatalog with nested service structures. This is the difference between a generic "roofing contractor" signal and a specific "asphalt shingle + metal roof + emergency repair" signal. When an LLM needs to match a homeowner's specific problem to a contractor's expertise, specificity wins.

Install schema using Google's Structured Data Markup Helper (free) or directly in your theme if you're on WordPress. Minimum required:

• LocalBusiness schema on homepage: company name, address, phone, service areas (cities + radius), hours, reviews.
• HasOfferCatalog schema on homepage and service pages: nested list of every service type and material option.
• Service schema on each service page: service name, description, service area, price range if available.
• AggregateRating on homepage and service pages: average rating + number of reviews.

Bonus: Add an "AI Summary" Block

After your schema is set, add a plain-text summary section on your homepage. This gives the AI a clean summary to digest:

Company Summary:
[Your Company Name] is a licensed roofing contractor established in [Year],
License #[#], based in [City, State]. We provide asphalt shingle installation,
metal roofing, flat roof repair, emergency storm damage assessment, and gutter
installation throughout a 25-mile radius. We've completed 500+ residential and
commercial roofing projects. Specialties: architectural shingles, torch-down
repairs, impact-resistant materials, seamless gutters. Service areas: [City],
[City], [City]. Average rating: 4.9/5 from 200+ verified reviews.

Why this works: when an LLM is parsing your page, this summary is a pre-formatted extraction point. Instead of the AI digging through your service pages to piece together what you do, it has a ready-made signal. This increases citation accuracy by 30 to 40% in our testing.

Week 3 to 4: Rewrite service pages. Don't stuff keywords. Write as if explaining to a homeowner and an AI simultaneously. Example structure:

[Service Name]: [Specific Definition]
We [specific process]. This works because [mechanism].
Service area: [specific cities and radius].
Why choose us for [service]: [2-3 differentiators with proof].
Recent work: [2-3 project examples with outcomes].
FAQ: [4-5 questions AI might surface in a response].

Minimum 600 words per service page. Semantic, clear, specific.

Month 2: Citations and Review Consolidation

Week 1 to 2: Audit your citations. Pull a spreadsheet with your company name, phone, address, website. Now check:

• Google My Business (correct? all services listed? service areas defined?)
• Yelp (complete profile?)
• Thumbtack, HomeAdvisor, Angi (if you're there, are they consistent?)
• Local contractor boards
• BBB

Fix NAP inconsistencies. Update phone number everywhere if it changed. Add missing service areas.

Week 3 to 4: Consolidate review signals. You have 200 reviews. Where are they?

Request reviews from recent clients specifically mentioning what service you did. "We had [roofing type] installed by [company name]" is stronger than "Great service!"

Target 20 to 30 new reviews with service-specific language. It takes time, but it compounds AI signal.

Month 3: Content and Entity Reinforcement

Week 1 to 2: Add a FAQ page. Not for Google. For LLMs. When an AI is asked "Do I need a new roof or can it be repaired?", it should find your answer. Write 10 to 15 FAQs that answer homeowner questions while establishing your expertise and service specifics. If you're in storm-prone markets, the 90-day storm content plan covers the exact FAQ structure that gets cited during demand spikes.

Week 3 to 4: Create a "Recent Projects" or case study section. 5 to 10 short case studies, each describing a specific roof type, problem, and solution. This creates entity reinforcement: this company has handled shingle roofs, flat roofs, storm damage, and so on.

The Real Play: From SEO Visibility to AI Visibility

You've spent years optimizing for a search engine that younger homeowners increasingly don't use first. Gen Z and young millennials ask ChatGPT "best roofer near me" before they ask Google.

The 2-year-old company isn't smarter than you. They're just shaped differently. And the good news? Your 200 reviews, your domain age, and your business stability are better signals to an LLM than they are to Google. You're starting this game with more firepower than you think.

You don't need a rebrand. You need a retrofit. And you don't need a consultant. You need 60 days of focused structure.

Start by running the free $0 AI audit to see exactly where you stand in ChatGPT and Perplexity today. It takes 15 minutes and tells you what to fix first.

Also worth checking: one issue that silently blocks AI crawlers on many roofing sites is a misconfigured robots.txt file. If GPTBot or PerplexityBot is disallowed, none of the above work matters until that's fixed.

Ready to audit your AI visibility?

Get your AI Visibility Audit and find out exactly which of your roofing pages get cited by LLMs, where you're losing to competitors, and your specific quick wins for AI search.