SOCi's 2026 Local Visibility Index tested thousands of business locations across ChatGPT, Perplexity, and Gemini. The results were stark. Google's local 3-pack shows 35.9% of tested businesses. Gemini recommends 11%. Perplexity manages 7.4%. ChatGPT recommends just 1.2%.

That means for every 100 trades businesses in a given area, ChatGPT surfaces about one. The other 99 are invisible when a homeowner asks "who should I call for a burst pipe" or "best electrician near me."

The businesses that break through are not the ones with the most reviews or the longest history. They are the ones that satisfy the six signals AI search engines check before recommending anyone. Get all six right and the recommendation rate climbs. Miss any one and the AI skips to the next candidate.

Signal 1: Data accuracy across every platform

SOCi found that business profile accuracy on ChatGPT and Perplexity sits at 68%. On Gemini, it is 100%. Gemini pulls local data directly from Google Maps, which means a business with a complete and accurate Google Business Profile is already in better shape on Gemini than on ChatGPT.

The gap matters because AI engines cross-reference information across Google, Yelp, Bing, Facebook, and the business website. When the hours on Yelp say 8 AM but the website says 7 AM, or the phone number on Google does not match the one on Bing, the AI loses confidence. It does not rank the business lower. It drops it from consideration.

Fixing this costs nothing. Audit the business name, address, and phone number across every directory where the business appears. Match them exactly. NAP consistency is the baseline. Without it, the other five signals do not matter.

Signal 2: Star rating above the AI floor

ChatGPT applies a star-rating floor of 4.30 for business recommendations, according to f9 Analytics. Perplexity's floor is 4.10. Gemini's is 3.90. Google Maps still surfaces businesses at 3.50 to 4.00, which means the bar for AI recommendation is higher than the bar for Map Pack visibility.

A contractor with a 4.1 average on Google may appear in the local pack but get skipped by ChatGPT because the rating falls below the 4.30 threshold. This is not a ranking penalty. It is an exclusion. The AI treats a 4.1 as too risky to recommend when a 4.5 option exists nearby.

Star ratings are the gatekeeper for AI visibility. A trades business below 4.3 needs recent positive reviews from real customers to cross the threshold. Volume matters less than the average and the recency of the most recent reviews.

Signal 3: Review response rate

SOCi describes review response rate as a "particularly high-leverage signal" because it shows the business is active and engaged. A 100% response rate on both positive and negative reviews signals that someone is running the shop. A 5% response rate signals that nobody is home.

AI engines treat engagement as a trust signal. A contractor who responds to reviews, answers questions on their Google Business Profile, and posts updates is sending the signal that the business is operational. A contractor who claimed their listing once and never touched it again looks abandoned.

This is the signal most trades businesses ignore. They have 200 Google reviews and have responded to three of them. Fixing it takes an hour a week. Reply to every review, thank customers for positive ones, and address concerns in negative ones. The AI reads this activity as proof the business is alive.

Signal 4: Content that answers specific questions

Research analyzing over 129,000 ChatGPT citations found that listicle and comparative content accounts for 25.37% of all AI citations. Definitive, structured content with clear headings, FAQ sections, and direct answers performs better than narrative paragraphs.

Content updated within the last 30 days receives 3.2 times more citations than older material. AI platforms, especially Perplexity, treat recency as a proxy for accuracy. A service page last updated in 2023 looks stale. The same page updated last month looks current and relevant.

For trades businesses, this means publishing pages that answer the specific questions homeowners ask. "How much does a water heater replacement cost in Bergen County" is a page AI can cite. "Quality residential solutions for Northern New Jersey" is not. The format matters. Answer-shaped content with clear headings and concise answers is what answer engine optimization is built around.

Signal 5: Structured data on the website

LocalBusiness schema gives AI engines a clean, machine-readable data source for the business name, address, phone number, hours, services, and service area. Without it, the AI has to parse the website and guess. With it, the AI reads the structured data directly and cross-references it against directory listings.

Google's own documentation recommends LocalBusiness schema for local businesses. A 2026 study by Evergrow Marketing tested schema markup over 10 weeks and found it did not directly affect search rank or Google Maps rank. But the study also noted that schema makes business data easier for AI engines to extract and confirm, which is the mechanism that drives AI recommendations.

Adding LocalBusiness schema takes about 30 minutes for a trades website. The payoff is not in Google rank. It is in AI confidence. A business the AI can read structurally is a business the AI can recommend with certainty.

Signal 6: Cross-platform authority and mentions

AI engines weight third-party mentions heavily because they are harder to fake than self-published content. A contractor mentioned in a local newspaper, a trade association directory, or a supplier's website gets a signal that a contractor with 800 Google reviews but zero press coverage does not.

This is why national franchises dominate AI recommendations. Roto-Rooter has consistent data across 50 directories, a Wikipedia page, and decades of press coverage. An independent plumber with a Google Business Profile and a Yelp page but no other presence is hard for AI to confirm. The 5WPR study found that 87% of independent HVAC and plumbing contractors have effectively zero AI citation share. National franchises pulled 19% of all consumer-intent citations between three brands alone.

Building cross-platform authority does not happen overnight. But claiming industry directory listings (PHCC for plumbing, ACCA for HVAC), getting listed on supplier websites, and earning local press coverage are all achievable for a one-to-five-truck operation. Each mention in a source the AI trusts adds a brick to the foundation.

How the six signals work together

No single signal gets a business recommended. AI search engines require all six to line up before they will put a business in front of a homeowner. A contractor with a 4.8 star rating, 500 reviews, and a perfect Google Business Profile but inconsistent hours on Yelp and no LocalBusiness schema is still missing two signals. The AI sees the gap and moves on to a business with fewer reviews but cleaner data.

The businesses that appear in AI answers are not the biggest or the highest-rated. They are the most confirmable. Every signal checks out across every platform the AI searches. That consistency is what makes the AI confident enough to say "call this one."

FAQ

What are the main signals AI search engines use to recommend local businesses?

The six signals are data accuracy across platforms, star rating above the AI floor (4.30 for ChatGPT, 4.10 for Perplexity, 3.90 for Gemini), review response rate, content that directly answers specific questions, LocalBusiness schema on the website, and cross-platform authority from third-party mentions. All six need to be present for AI engines to recommend a business with confidence.

Why does ChatGPT recommend fewer businesses than Gemini?

ChatGPT recommends only 1.2% of tested business locations, compared to 11% for Gemini, according to SOCi's 2026 Local Visibility Index. Gemini pulls local data directly from Google Maps, where business profile accuracy is 100%. ChatGPT and Perplexity have 68% profile accuracy, which means they exclude businesses they cannot verify with high confidence rather than ranking them lower.

How often should a trades business update its website for AI search?

Content updated within the last 30 days receives 3.2 times more AI citations than older material, based on analysis of 129,000 ChatGPT citations. Trades businesses should update their Google Business Profile weekly, add new photos quarterly, and refresh service pages when offerings or pricing change. Freshness signals that the business is active.

Does review response rate actually affect AI visibility?

Yes. SOCi identifies response rate as a high-leverage signal because it demonstrates active engagement. Businesses with a 100% review response rate signal that someone is operating the business. AI engines treat low response rates as a sign the business may be inactive, which reduces recommendation frequency even when the star rating is high.

What is the fastest way to improve AI search visibility for a trades business?

Audit and fix NAP consistency across all directories (free, 30 minutes), claim and complete a Bing Places listing (free, 15 minutes), add LocalBusiness schema to the website (free, 30 minutes), and start responding to every review (ongoing, 1 hour per week). These four actions address the most common reasons trades businesses are invisible to AI: inconsistent data, missing directory presence, missing structured data, and low engagement signals.

Run your free AI visibility audit to see which of the six signals your trades business is missing. The audit checks 12 data points across ChatGPT, Gemini, and Perplexity and flags exactly what to fix to start appearing in AI answers.