Tony Mannino owns a small excavation outfit in Jackson, NJ. Six trucks, eight guys in the field, mostly residential foundation digs and septic work. Last March, a homeowner in Howell told him: "I found you on Perplexity. It said you do same-week septic installations and that’s exactly what I needed."
Tony had to look up what Perplexity was.
That call closed at $8,400. It’s the first lead Mannino can attribute directly to AI search. It probably won’t be the last — the same homeowner pattern is now showing up across our client base, especially for higher-ticket residential trades where homeowners spend a few weeks researching before they pick a contractor.
Here’s the story of what made Tony’s site the one Perplexity cited — and what every contractor can learn from a single page on his site that did most of the work.
What the homeowner actually asked
The homeowner shared the Perplexity conversation. The query was: "Who can do a residential septic system replacement in Howell NJ within a week? Need fast turnaround."
Perplexity returned three contractors. Mannino was first. The cited reason: a service page on his site that explicitly stated turnaround windows for septic installations, with named townships and price ranges.
That page didn’t exist six months earlier. We built it for him as part of his Genesis launch.
The page that did the work
The page is /services/septic-system-replacement/ on Mannino’s site. Nothing fancy. About 1,200 words. But it’s structured in a very specific way that AI engines reward:
Direct, specific opening
The first paragraph names the service, the townships served, the typical timeframe, and the typical cost range. No marketing fluff. No "with over 15 years of trusted experience..." setup. Just facts an AI engine can extract and quote:
Mannino Excavation handles residential septic system replacement across Jackson, Howell, Manchester, Brick, and Toms River, NJ. Standard turnaround for a full septic replacement is 5–10 business days from approval to working system, depending on permit timing. Typical cost ranges from $9,500 to $18,000 for a 1,000-gallon system on a standard residential lot.
Compare this to the typical contractor service page: "We pride ourselves on quality septic installations done right the first time. Our experienced team has been serving New Jersey families for decades." The first one is citable. The second one is unprintable.
FAQ structure with H2/H3 hierarchy
The body of the page is a series of H2 questions with paragraph answers directly underneath. Verbatim from the page:
- How fast can you replace a residential septic system?
- How much does septic replacement cost in Ocean County?
- Do you handle the township permits?
- What happens to my old septic tank?
- Will you tear up my yard?
- Do you offer financing for septic replacement?
Each of those is a real question homeowners ask AI search engines word-for-word. The page answers each one in 80–150 words with specific facts. That’s the citation-friendly format.
Specific named entities
The page mentions Tony by name. It mentions the town of Jackson, NJ specifically. It names the brand of tank Mannino installs (Infiltrator), the township permit process they handle, and the typical excavation depth (4–6 feet). Specific named entities give AI engines something to anchor on. Vague generic content gets ignored.
Pricing ranges, not "call for pricing"
This is the one most contractors fight us on. They don’t want to publish prices. We get it. But AI engines retrieve and cite content based on whether it answers the user’s actual question — and "how much does it cost" is the most-asked contractor question on every search platform. A page that doesn’t answer that question won’t get cited. A page that says "$9,500–$18,000 depending on tank size and lot conditions, free site visit to confirm" gets cited and gets clicks.
The supporting infrastructure that helped
The septic page didn’t do the work alone. A few other things made Mannino visible to Perplexity:
llms.txt at the site root
Mannino has a 60-line markdown file at buildatlas.ai/mannino/llms.txt that tells AI models exactly who he is, what he does, what towns he serves, and what specific services come with what specific pricing. It’s the AI-search equivalent of a robots.txt file — not yet required, but signals you understand the channel.
FAQ schema markup on every service page
Every FAQ section on Mannino’s site is wrapped in JSON-LD FAQPage schema. This tells search engines (including AI-search engines that crawl the same indices) that the content is structured Q&A. It’s a small but real ranking signal.
Real Google Reviews with location tags
Mannino has 47 Google Reviews averaging 4.9 stars. Many of them mention specific townships ("Tony’s crew did our septic in Howell — on time and clean") which AI engines pick up as location-specific trust signals.
A clean Google Business Profile
Service area set correctly. Hours posted. Photos posted weekly. The basics. AI engines pull from the same Google Business data that powers Map Pack rankings, so this work doubles up.
Could you replicate this?
Yes. None of it is technical wizardry. The recipe:
- Pick your three highest-revenue services. Build a dedicated page for each one.
- Open each page with a one-paragraph factual summary — service, area, timeframe, price range. No marketing voice.
- Structure the body as 5–8 H2 questions homeowners actually ask, with paragraph answers underneath.
- Name your towns, your tools, your team. Specificity gets cited; vagueness gets ignored.
- Publish pricing ranges, not "call for pricing." Even rough ranges work.
- Add FAQ schema markup. Add an llms.txt at site root.
- Keep your Google Business Profile updated weekly.
That’s the whole playbook. Tony’s site is the proof it works — one well-structured service page, one Perplexity citation, one $8,400 job that wouldn’t have existed otherwise.
What this means for your shop
AI search is currently 10–25% of contractor-finding traffic. By 2027 it’ll be 40%+. The contractors visible to ChatGPT, Perplexity, and Google AI Overviews now will compound that visibility into next year. The ones who wait will pay more in ads to compensate for the leads they don’t see.
The work is straightforward. The cost is mostly time. And one well-cited page can pay for itself in a single closed job.
If your current site is full of "trusted experience you can count on" boilerplate, you’re invisible to AI. Atlas Genesis builds contractor sites with the structure AI engines reward by default — FAQ-first service pages, named entities, transparent pricing ranges, llms.txt, FAQ schema, and a Google Business Profile we keep updated. Same shape that put Tony’s septic page on Perplexity. Same shape that puts your business in the answer when homeowners ask.