AI Search Recommends Just 1.2 Percent of Local Contractors. Here Is the Trust Stack That Decides Who Makes the List.

SOCi just published a number that should ruin somebody's Saturday morning: only 1.2 percent of local business locations ever get recommended by AI search. Out of 350,000-plus locations analyzed across 2,751 brands. One in eighty.

If that does not make you sit up, the next number will. Whitespark's Q2 2026 local data set found that AI Overviews now appear in 68 percent of local searches overall, 92 percent of informational local queries, and 97 percent of hybrid intent queries like "average cost of HVAC replacement in [city]." Local map packs only show up for 39 percent.

Translation for the contractor reading this: the spot you used to fight for at the top of Google is gone for most of the questions your customers are asking. The new spot is inside the AI summary. And the AI is choosing one or two contractors from your entire metro to mention by name.

How the AI actually picks

This is not Google ranking. It is RAG — retrieval-augmented generation. The AI takes a customer query like "best plumber for tankless install in Greenville," fans it out into half a dozen sub-queries, pulls candidates from independent sources, and synthesizes a two-or-three-business recommendation.

The contractor it picks is the one that shows up consistently across the most independent sources with the most corroborating signals. If your name appears on your own website and basically nowhere else, you do not exist for that query. End of story.

The five layers of the AI trust stack

Every name you see cited in ChatGPT, Gemini, Perplexity, or Google AI Overviews ran the same gauntlet:

• Entity verification. Identical NAP across 10-plus platforms — Google Business Profile, Yelp, Facebook, Bing Places, Apple Maps, BBB, Foursquare, Nextdoor, Angi, plus your industry-specific directories. One mismatch is a deduction.
• Service-geography match. Your content explicitly names the services you offer and the geographies you serve, in the same language customers use to ask the AI.
• Social proof weight. Review volume, recency, sentiment, and response behavior. Active profiles on review platforms drive a documented 3x citation boost (SE Ranking).
• Third-party mentions. Local news, podcast appearances, industry publication features, guest articles. The AI uses these as evidence you are an established entity, not just a website.
• Content freshness. Posts in the last 30 days, new photos, review responses, updated service descriptions. Stale profiles get skipped.

Miss any one of these layers and you drop into the 98.8 percent that AI never names.

The schema number that nobody is talking about

Semrush ran the math. Properly implemented LocalBusiness schema = 45 percent higher AI citation rate. FogLift's 2026 study found that pages with FAQ schema are 2.8x more likely to be cited in AI answers than pages without it.

Almost no contractor website has either, properly. Most have a generic theme that ships with broken or partial schema. That is not optimization. That is a leaky bucket. The contractors who fix this in the next 60 days are about to be cited at three times the rate of their competitors who do not, for zero additional ad spend.

The 30-day move

Day 1-3. Run a citation scan with BrightLocal, Moz Local, or Whitespark. You will find 20 to 40 listings with outdated information, old phone numbers, prior addresses, and legacy business names. Document the corrections.

Day 4-7. Standardize the master record. Pick the exact business name, exact address format, primary phone. That is the source of truth from now on.

Day 8-14. Fix the top 10 directories manually. Then push corrections to the data aggregators — Data Axle, Localeze, Foursquare, Factual. Aggregators feed hundreds of downstream listings automatically.

Day 15-21. Add LocalBusiness schema with AggregateRating, Service schema for every service page, FAQ schema for the questions you actually get asked, and Review schema where you have permission to display customer quotes. Validate everything in Schema.org's tool.

Day 22-30. Run a Share of Model audit. Build a list of 25 queries your customers might ask an AI — "near me" queries, "best in [city]" queries, problem queries, pricing queries. Run each manually in ChatGPT, Gemini, Perplexity, and Google AI Mode. Log whether you appear, who else appears, and which sources got cited.

Most local contractors start under 10 percent Share of Model. The goal in 90 days is 30 to 40 percent.

Why this window is closing

Every month you wait, the contractors who started this in January 2026 accumulate more reviews, more third-party mentions, more schema completeness, and more citation history. AI systems get progressively more confident in their existing recommendations and progressively less likely to swap them out for a newcomer.

The 1.2 percent that AI cites today is not a fixed list. But it is hardening fast. The contractors who own the citations in 2026 are going to own the recommendations through 2028 and beyond, the way local pack winners owned Google through the 2010s.

The customer is no longer choosing between three quotes. They are taking the one or two names the AI gave them. Be one of those names. Or be invisible. There is no third lane.