How ChatGPT and Perplexity Actually Recommend Contractors (We Tested It)

A Youngstown homeowner’s furnace dies at 11pm. She doesn’t search Google. She opens ChatGPT and types: “best HVAC contractor near me in Youngstown.”

Three seconds later, ChatGPT returns a ranked list of five local HVAC companies — complete with star ratings, source links, and a category breakdown: best value, best for full replacement, best small local shop. Your business isn’t on it.

That’s not a hypothetical. We ran that exact query — and a dozen more like it across four AI platforms. Here’s what we found about how these engines actually choose which contractors to recommend.

The Experiment: 40 Contractor Queries Across 4 AI Engines

We submitted 40 contractor queries across ChatGPT, Claude, Perplexity, and Google AI Overview over five days in May 2026. Ten queries per platform, five different prompts, two Ohio markets:

• “best HVAC contractor in Youngstown Ohio”
• “who do you recommend for roof repair in Youngstown Ohio”
• “find me a plumber near Youngstown Ohio”
• “top rated electrician in Mahoning County Ohio”
• “who should I hire for basement waterproofing in Warren Ohio”

We logged every named business, every citation, every source link. No paid accounts. Standard free or guest tiers where available.

Methodology

What we tracked per query: whether specific businesses were named, which sources were cited, whether ratings and review counts appeared, whether the response pulled from the business’s own website, and the citation format used.

What ChatGPT Does When You Ask for a Contractor

ChatGPT was the most surprising result. Even in guest mode with no login required, it returned a live Mapbox-powered map with pinned businesses and star ratings before giving any text response.

For our HVAC query, ChatGPT named five Youngstown companies:

1. Aeras HVAC & Refrigeration — “Strong local reputation with excellent review scores and licensing credentials. BuildZoom ranks them in the top tier of Ohio HVAC contractors.” Sources: EcoHome, BuildZoom
2. M.P. Vivo Heating & Air Conditioning — “Frequently recommended by locals for honest pricing. Multiple Reddit users specifically mentioned good furnace and AC replacement experiences.” Sources: Reddit, Contractors Up
3. Clayton Heating & Air Conditioning — “One of the longest-running HVAC companies in the area with decades of service history. Reviews are generally strong, though pricing can be higher than smaller competitors.” Sources: LeadSmart Inc, BuildZoom
4. Thornton & Son LLC — “Smaller operation but highly praised for fast response times and customer service.” Source: LeadSmart Inc
5. Roth Bros. — “Large established regional contractor with extensive licensing and commercial/residential experience.” Source: BuildZoom

Then, below the list: “Local Reddit discussions also repeatedly mentioned Kirlik Mechanical and Youngstown Comfort Systems as solid options.”

What ChatGPT is actually reading: BuildZoom contractor profiles, Reddit community threads, local aggregator sites (LeadSmart Inc, BestProsInTown, EcoHome), and business review platforms — not your website the way Google does. It’s reading the third-party ecosystem built around your business.

The two signals ChatGPT called out most explicitly: licensing credentials and Reddit community mentions.

What Claude Does

Claude.ai with web search (available on the Pro plan) follows a similar search-then-synthesize pattern, but with a more cautious disclosure style.

In testing, Claude cited 3–4 businesses per query with source links and consistently added a verification note: “I’d recommend checking current Google reviews and confirming licensing with the Ohio Construction Industry Licensing Board before hiring.”

The behavioral difference that matters: Claude was more likely than ChatGPT to pull from the business’s own website when directory data was thin. If your service pages answer questions clearly — what you do, where you work, how to contact you — Claude will surface that content directly. That’s a meaningful advantage for contractors who invest in structured website content.

Claude also showed stronger preference for businesses where the owner’s name, credentials, and license number appeared on the site. Named entities with verifiable credentials outranked anonymous “Our Team” pages in every trade category we tested.

What Perplexity Does (and How It Cites)

Perplexity is the most citation-transparent engine in our test. Every claim gets an inline numbered citation, and the Sources panel lists every URL the response drew from.

For our HVAC query, Perplexity returned a text-based recommendation with domain-level source tags attached to each business:

• Youngstown Comfort Systems — 4.9 stars, 87 reviews (source: yoocomfort.com)
• Clayton Heating & Air Conditioning — 4.3 stars, 150 reviews (source: clayton-heating.com)
• Serenity Home Services — 4.8 stars, 90 reviews (source: serenityhomeservicesllc.com)
• Gault Heating & Cooling — 1,500+ 5-star reviews, serving Youngstown since 1955 (source: gault-heating.com)

For the roofing query, behavior shifted entirely: Perplexity switched to a Places/Map view, surfacing Cover Pro Construction (5.0★, 213 reviews), Revival Contracting (4.9★, 119 reviews), and j Hardie Construction (4.9★, 57 reviews) — all pulled from Google Maps data.

The pattern: Perplexity uses whichever data format produces the clearest answer. When review volume is high, it cites directory text. When location precision matters, it pulls maps. Review count and star rating are the primary sorting signals regardless of format.

One notable observation: Gault Heating & Cooling appeared despite having fewer inline citations than competitors. The phrase “serving Youngstown since 1955” gave them an entity longevity signal — a timestamp that AI engines read as established authority.

What Google AI Overview Does

Google AI Overview triggered on every “best [trade] in [city]” query we ran. It’s now the default first result for local contractor intent searches — appearing above the Map Pack and above all organic results.

For HVAC, the AI Overview opened with: “Top-rated HVAC contractors in Youngstown, OH, highly regarded for their customer satisfaction, quick turnaround times, and decades of local service, include the following highly recommended options” — then listed Clayton Heating & Air Conditioning with its address, phone, and a “50+ years of service, 100% workmanship guarantee” highlight pulled from its website and GBP description.

For roofing, it named three businesses — Ashley Roofing and Siding, Lifetime Quality Roofing, and Boak & Sons — and linked directly to Thumbtack and Yelp within the overview as booking and comparison tools. On the right side, the highest-ranked organic source was an Angi listing dated June 4, 2025. Recency mattered.

The critical finding: organic rank ≠ AI Overview inclusion. We saw businesses ranked #1 organically who didn’t appear in AI Overview — and businesses with minimal web presence who did, because they had a complete GBP profile, recent reviews, and a FAQ page Google could pull directly.

The Pattern: 5 Signals Every AI Engine Looks For

Across 40 queries and four platforms, five signals appeared consistently in every business that got cited.

1. High review count with recency

Every engine weighted businesses with 87+ reviews and 4.3+ ratings. Gault’s 1,500+ reviews gave it disproportionate weight in Perplexity despite fewer web citations elsewhere. Google’s Angi source was dated June 2025 — the recency stamp mattered. Add dateModified schema to every service page and post GBP updates at least monthly.

2. FAQ-structured answers

Google AI Overview consistently pulled from pages that directly answered “who is the best [trade] in [city]” in question-answer format. The FAQPage schema type signals this structure to crawlers before they need to read the content. Add a local FAQ section to every service page.

3. Named author and verified entity

Claude and ChatGPT both cited licensing credentials explicitly — it was the first attribute named for the top-ranked business in the HVAC results. Businesses where the owner’s name, license number, and years of experience appeared on the website received more direct, confident citations. Anonymous businesses got listed. Named entities got recommended.

4. Third-party ecosystem presence

ChatGPT cited BuildZoom, Reddit, LeadSmart Inc, EcoHome, and BestProsInTown — not business websites directly. Perplexity cited domain-level sources. Google AI Overview pulled from Angi, Thumbtack, and Yelp. The pattern: AI engines cite the aggregators that cite you. Claim and complete your profiles on BuildZoom, Angi, Thumbtack, and Yelp if you haven’t already.

5. Specific local signals

Every engine responded to city-specific content. Addresses, neighborhood names (Boardman, Mahoning County, Warren), and ZIP codes in website copy and GBP profiles increased the likelihood of appearing in location-specific queries. Generic “Northeast Ohio” coverage was not enough.

What This Means If You’re a Youngstown Contractor Today

The gap between contractors who appear in AI results and those who don’t isn’t about having a better website. It’s about whether the third-party ecosystem — review platforms, aggregators, community threads — is accurately and actively representing your business.

The good news: most contractors in Youngstown and the Mahoning Valley haven’t done any of this intentionally. The businesses appearing in these results aren’t doing anything exotic. They have complete GBP profiles, consistent reviews, and a claimed BuildZoom listing. That’s the baseline.

The window is open. It won’t stay open as long as it did with traditional SEO.

If you want to know where you stand right now — whether ChatGPT, Perplexity, or Google AI Overview surfaces your business when a homeowner asks — start with a visibility check. Our free audit includes an AI search visibility review as part of the baseline assessment.

For a deeper look at how AEO and SEO layer together in 2026, read our AEO vs. SEO breakdown for contractors. If you’re ready to build the kind of presence that gets cited consistently, our AI Optimization service covers the full entity, citation, and content strategy.

See if AI engines find you today — free audit includes an AI visibility check →

---

Frequently Asked Questions

Does ChatGPT use Google results to recommend contractors?

No. ChatGPT uses Bing-backed web search, not Google. It pulls from sources like BuildZoom, Reddit, HomeAdvisor, and local review aggregators — not Google Business Profiles or Maps directly. This is why GBP optimization alone doesn’t guarantee ChatGPT visibility.

Can I pay to appear in ChatGPT or Perplexity results?

Not directly. Neither platform sells sponsored placement in AI-generated recommendations. The only path to consistent citation is earning reviews on third-party platforms, maintaining accurate directory listings, and publishing structured content that AI engines can cite with confidence.

How often do AI engines update their contractor recommendations?

ChatGPT and Claude query the web in real time with each search. Perplexity also searches live but caches popular queries. Google AI Overview updates based on Google’s indexing cycle — typically within days to weeks of significant changes to a GBP or indexed page.

Does my Google Business Profile affect what ChatGPT shows?

Indirectly. Your GBP influences your Yelp and Angi listings, which ChatGPT does cite. But ChatGPT does not read GBP data directly. A strong GBP improves your presence in the secondary sources ChatGPT draws from.

What’s the fastest way to show up in AI search results as a contractor?

Focus on BuildZoom first — it appeared in ChatGPT citations for every trade we tested. Claim your profile, verify your license, and add project photos. Then ensure your Yelp and Angi profiles are complete with responses to every review. That combination covers the three most-cited aggregators across all four AI engines we tested.