How Local Service Businesses Get Cited in AI Overviews: A City-by-City Playbook for 2026

Key takeaways

• AI Overviews now appear in an average of 68% of local business queries across plumbers, lawyers, dentists, optometrists, medical clinics, and real estate agents, per Whitespark's 540-query study across Houston, Phoenix, and Denver. Prevalence is consistent by industry regardless of city, not by location.
• AI local packs surface only about a third of the businesses a standard 3-pack does (5,943 vs 18,330 unique businesses in Joy Hawkins' Sterling Sky tracking), and they drop call buttons entirely.
• Yelp alone accounted for 512,680 AI citations in a Foundation Inc./AirOps analysis of 28 million responses, more than BBB, Angi, Thumbtack, HomeAdvisor, and Nextdoor combined, and over half came from "near me" queries.
• Review volume needed to compete varies wildly by vertical: plumbers need a median of 215 reviews, dentists 346, home health care businesses just 20, per Local Falcon's analysis of 50 million search results.
• Mid-authority domains (DR 46-75) earned 46% of ChatGPT citations in August 2026, while the top DR 91-100 tier collapsed to about 3%, so a modest local-business website genuinely competes.

Why this city doesn't care about your 3-pack ranking anymore

I'll say the uncomfortable part first: if you're a plumber, dentist, or personal injury lawyer and you're still obsessing over your position in the Google 3-pack, you're optimizing for a shrinking slice of the pie. Whitespark ran a 540-query study across six local-business verticals in Houston, Phoenix, and Denver and found AI Overviews showing up in an average of 68% of those searches. Local Falcon's broader 60,000-query study, which included a lot of navigational and Boolean queries where AI rarely appears, put the number at 40.2%. Both are right. The honest range is 40-68% depending on how commercial the query is, and for the verticals that pay the bills, you're at the high end.

Here's the split that actually matters for planning: pure local-intent queries like "pizza delivery near me" still trigger the map pack 93% of the time and AI Overviews only 15%. But anything with an informational or hybrid dimension, like "best dentist for invisalign Phoenix" or "emergency plumber cost Denver," flips that ratio hard. Informational queries trigger AI Overviews 92% of the time. Hybrid queries hit 97%. If your service business lives or dies on "best [service] in [city]" searches, you are now mostly being evaluated by an AI, not a ranking algorithm sorting ten blue links.

And the AI local pack, when it does appear, is a much smaller room. Joy Hawkins at Sterling Sky tracked AI local packs surfacing 5,943 unique businesses against 18,330 in standard 3-packs across the same query set, roughly 32% coverage, and 88% of the 322 markets she examined showed fewer unique businesses in AI packs than in 3-packs. AI packs typically show one or two businesses instead of three, and they drop the call button. That's not a tweak. That's a different game with a different roster of winners.

The industry, not the city, sets your odds

The most useful thing in Whitespark's data is the city-by-city breakdown, because it reveals something counterintuitive: your industry matters far more than your city.

Legal queries sit around 70% AI Overview exposure in all three metros. Real estate sits highest of any vertical, with Denver's local pack dropping as low as 17% for real estate searches. No correlation showed up between local-pack prevalence and AI Overview prevalence in the dataset. So if you're running a multi-location plumbing or dental brand across Houston, Phoenix, and Denver, you can reasonably expect similar AI exposure in each market, which means the same optimization playbook should transfer between your locations, with adjustments for smaller towns where the pattern may shift.

Where the AI is actually getting its answers from

This is the part most local businesses get wrong. They assume ranking well in Google organic search is the whole job. Per Whitespark's guide to AI Mode for local businesses, being in the top 10 of organic results gives you only a 25% chance of appearing in AI Overviews. Organic rank is a prerequisite, not a guarantee.

So where is the AI actually pulling its recommendations from? Local Falcon's 2026 analysis found that 60% of AI Overview citations in local-business queries point to third-party publishers, not the business's own website. Yelp is the dominant one. Foundation Inc. and AirOps analyzed 28 million AI responses to small-business queries across ChatGPT, Gemini, Perplexity, and Google AI Mode and found Yelp pulling 512,680 citations, more than Angi, BBB, Thumbtack, HomeAdvisor, and Nextdoor combined, a 3.4x lead over the next closest platform. On Google AI Mode specifically, Yelp captured 72.5% of citations within that competitive set. More than half of Yelp's citations came from "near me" queries, and Yelp's citation volume grew 19x between September and November 2025 while the category overall grew 3x. Cleaning and plumbing "near me" searches alone generated over 2,100 Yelp-cited answers each on Perplexity and Google AI Mode combined.

I'll flag the counterpoint honestly, because the data isn't unanimous here: a separate Yext study using a different methodology found 86% of AI citations come from sources brands already control, with Gemini citing brand-owned sites 52.2% of the time versus ChatGPT's 48.7%. That's a meaningfully different conclusion from the 60/40 third-party split. My read: the platform you're optimizing for changes the math, and ChatGPT in particular leans more third-party than Gemini does. Either way, a Yelp profile with current hours, fresh photos, and a steady reply cadence isn't optional anymore. It's the front door.

One caveat worth knowing before you pour effort into Reddit threads as a citation strategy: Promptwatch's data shows Reddit's share of ChatGPT citations collapsed from a steady 3.8% to under 1% on a single day in August 2026, an 86% relative drop. Google AI Overviews and AI Mode declined only gradually over the same window. If your local-citation strategy for ChatGPT specifically leans on forum mentions, that bet got a lot riskier overnight.

The review math nobody tells you by vertical

Here's where generic local SEO advice falls apart. "Get more reviews" means wildly different things depending on what you do. Local Falcon analyzed 50.4 million US search results across nearly 2,000 business categories and found the review volume needed to compete in the 3-pack ranges from almost nothing to nearly a thousand, depending on vertical.

The star-rating floor sits at 4.3-4.5 across almost every category, even though volume requirements swing by 50x. Most actual market leaders sit at 4.8-4.9 stars, not the bare minimum. And location matters less than you'd think: businesses in big cities need roughly twice as many reviews as rural competitors, not the 10x gap that gets repeated in local SEO circles.

Recency matters as much as volume. Whitespark's AI Mode guide treats reviews as one of the primary sources AI pulls from when assembling local-business recommendations, which means a dentist with 500 reviews and nothing new in four months looks stale to the algorithm even if the historical volume looks strong. Build review collection into your actual service workflow, not a quarterly push.

A city-by-city action checklist

Whitespark's guide to AI Mode lays out a practical sequence, and it's the closest thing to an operational playbook I've seen for this specific problem. I'd run it per market if you're multi-location, since AI Overview exposure is consistent by industry but third-party citation coverage (local news, chamber directories, "best of" lists) is genuinely local and has to be rebuilt city by city.

1. Audit NAP consistency across your website, Google Business Profile, and every third-party listing in that specific city. Conflicting phone numbers or addresses erode AI confidence fast.
2. Pull real customer FAQs from front-line staff, phrased the way customers actually type them, not the way marketers write headlines.
3. Manually run your own FAQ list through AI Mode and ChatGPT to see which sources get cited, and whether your business shows up on them.
4. Prompt the same tools with competitor brand names in that city to surface publications and directories you're missing.
5. Build unstructured citations specific to that market: local news mentions, sponsorship pages, chamber of commerce listings, job boards, community hubs. Whitespark notes Google AI Mode leans heavily on exactly this kind of citation for local search responses.
6. Push for inclusion on Yelp "10-Best" style roundups for that city, given Yelp's outsized citation share.
7. Photograph everything and publish broadly. AI Mode surfaces large image galleries, and GBP data backs this up: profiles with 100+ photos correlate with 520% more calls.
8. Enroll in Reserve with Google if it's cited frequently for your query set in that market.

One honest caveat from Whitespark that's worth repeating to clients: each AI Mode user can see different results for the identical query, which makes consistent tracking genuinely hard, and the AI can hallucinate by pulling from outdated or unrelated sources. This isn't a set-it-and-forget-it channel.

What content actually earns the citation

Content type matters more than people assume, and the data here is specific enough to act on. Promptwatch's citation-type research shows that for commercial-intent prompts, the kind closest to "best plumber near me" searches, ChatGPT cites product and landing pages, listicles, and comparisons for roughly 70% of citations combined. That means your service pages need actual pricing context and clear positioning, not vague brochure copy. For informational prompts like "how much does X cost," how-to guides and documentation punch above their average weight, which supports building out genuine cost-calculator and FAQ content rather than a single generic services page.

Don't assume you need a massive, famous website either. Promptwatch's August 2026 domain-rank data found mid-authority domains in the DR 46-75 range earned 46% of all ChatGPT citations, while the top DR 91-100 tier actually collapsed from about 7% to 3% over the month. A well-built local service site with a modest domain rating is competing in the real zone, not getting drowned out by national brands.

It's also worth setting realistic expectations about how crowded these answers are. Nearly half of ChatGPT responses with citations reference 10 or more different domains, according to Promptwatch's domains-per-response data. You're one of several sources in most unbranded local searches, not the only answer. Branded queries narrow that considerably, down to about 21% citing 10+ domains, which is one more argument for building a strong, ownable brand name in your market rather than competing purely on generic service terms.

Tools worth knowing for tracking this

You can't manage what you can't see, and AI Overview citations don't show up in a standard rank tracker. For brands and agencies managing this across multiple cities or franchise locations, a dedicated AI visibility platform like Promptwatch tracks citation sources, crawler activity, and prompt-level visibility across ChatGPT, Gemini, Perplexity, AI Overviews, and AI Mode, which matters once you're running the same playbook across several metros and need to know which city's citation-building is actually working.

Promptwatch

Track and optimize your brand's visibility in AI search engines

View more

For local-citation-specific work, a few tools are built closer to the ground-level task:

For businesses evaluating a broader set of AI visibility platforms beyond what's listed here, the directory at bestgeosoftware.com breaks down more options by feature set and price point.

What to actually do this quarter

Pick one city if you're multi-location, run the full checklist above end to end, measure what citation sources start appearing when you prompt AI Mode and ChatGPT with your own FAQ list, then replicate in the next market. Don't try to fix every city simultaneously; you'll spread review-collection and citation-building effort too thin to move either number.

The businesses getting cited in AI Overviews right now aren't necessarily the ones with the best website. They're the ones where a dozen independent sources, Yelp, local news, chamber directories, review platforms, agree on the same name, phone number, and service list without contradiction. That consistency is the actual ranking factor now, city by city.