A contractor's website has a homepage, a services page, a contact page, maybe a blog, and usually a locations page that lists every city served. That layout worked fine for traditional Google search. It is missing what AI engines actually cite.
New data from the Cheers Market Baseline, which tracked 74,969 AI citations to contractor websites across ChatGPT, Google AI Mode, Gemini, and Perplexity between August 31 and September 20, 2026, reveals exactly which pages AI engines reach for. The answer changes what a trades business should build first.
What AI engines actually cite
Of 74,969 citations to contractor websites, here is where they went:
- Homepage: 43.1%
- Service-plus-city page (a page naming both a service and a city): 17.1%
- Service page (no city in the path): 16.9%
- Location or city page (no service keyword): 14.0%
- Other: 3.2%
- Booking, offers, or financing: 2.3%
- Blog or resources: 1.5%
- About or contact: 1.2%
- Reviews or testimonials: 0.7%
Service, location, and service-plus-city pages combined took 48% of all citations. Blog posts took 1.5%. Review pages took 0.7%.
For trades businesses that have been told to "start a blog" for SEO, that is the wrong priority for AI search. The pages AI engines cite are the ones that name a specific service in a specific place.
The page type that matters most
A service-plus-city page is the highest-value page a contractor can build for AI visibility. It is a page like "/plumbing-repair-bergen-county-nj" or "/hvac-installation-passaic-county-nj" that names the service and the area in the URL, title, and heading.
Perplexity cited service-plus-city pages more than any other page type, at 28.6% of its contractor-website citations. Gemini cited them at 26.0%. ChatGPT at 17.4%.
Google AI Mode was the exception. It sent 66.9% of its website citations to homepages, probably because it pulls the website link from each business's Google Business Profile. The other three engines spread their citations across service and location pages.
For guidance on adding the structured data that helps AI engines read these pages, see how to add LocalBusiness schema to a trades website.
How the page type changes by trade
Plumbing answers cited homepages 35.7% of the time and service-plus-city pages 18.3%. HVAC citations went to homepages 45.8% and service-plus-city pages 15.4%. Electrical showed a similar pattern: 48.8% to homepages, 13.4% to service-plus-city pages.
Roofing leaned hardest on homepages at 67.7%, but that trade had only 2,631 cited URLs in the sample, so the number is less reliable.
The pattern across all trades is consistent: the homepage gets the most citations, and the service-plus-city page gets the second most. A contractor without a service-plus-city page is leaving the second most-cited page type empty.
What a service-plus-city page should include
Based on what AI engines cite and what makes a page useful to a model trying to answer "who is the best plumber in [city]," a service-plus-city page needs:
1. The service and the city in the page title and H1. Not "Services" or "What We Do." "Plumbing Repair in Bergen County, NJ" or "HVAC Installation in Passaic County, NJ." The page title is what AI engines read first.
2. A clear service description. What the business does, written the way a customer would describe the job. "Water heater replacement, drain cleaning, pipe repair, and emergency leak detection in Bergen County." Not "quality solutions for all your plumbing needs."
3. The service area, named specifically. Not "serving the tri-state area." List the actual towns. "Serving Hackensack, Englewood, Teaneck, Fort Lee, and surrounding Bergen County communities." Specificity helps AI match the page to a homeowner's question.
4. Contact information that matches every other source. The name, address, and phone number on the page must match the Google Business Profile, the BBB listing, the Angi profile, and every other directory. AI engines hedge when sources disagree.
5. Local proof. Real examples of work done in the area. Not fabricated testimonials. "Replaced a 50-gallon water heater in a Hackensack home in March 2026" is the kind of specific, verifiable detail that gives a model something to cite.
6. LocalBusiness schema markup. Structured data that tells AI engines the business name, address, phone, hours, and service area in a format machines can parse. It should match what the page shows.
For more on why specificity drives AI visibility, see content specificity and AI visibility for contractors.
What not to build
Do not build a page for every town with only the name swapped. Google's spam policies call that doorway abuse. Each page needs real local content, not a template with the city name find-and-replaced.
Do not rely on a blog for AI citations. Blog posts got 1.5% of citations. A blog can help with traditional SEO and brand authority, but it is not what AI engines reach for when a homeowner asks for a recommendation.
Do not bury service information on a generic services page. A page called "Services" that lists everything from drain cleaning to water heater installation gives AI engines nothing specific to cite. One page per service, with the city named, is the format that gets cited.
The trade-specific page a contractor needs first
A plumber should build one page per county served, titled and structured as "Plumbing Repair in [County], [State]." An HVAC contractor needs "HVAC Repair in [County], [State]" and "HVAC Installation in [County], [State]" as separate pages if both services are distinct enough to warrant their own URLs.
Start with the county where the business has the most customers. Build that page with real local details, verify the contact information matches across all directories, add LocalBusiness schema, and check whether AI engines cite it within two weeks.
For the broader picture of what ChatGPT reads when recommending trades businesses, see what ChatGPT cites when recommending trades businesses.
Ready to see where your business stands in AI search? Run your free AI visibility audit at FirstCall Visibility.