By Rahul Saini, Author at Search Counsel Co. Last updated [JULY] 2026.

You have probably read that AI Overviews appear on 68% of local searches. You have probably also read that they appear on 7%. Both numbers come from real studies, and both are close to useless on their own. The truth underneath them is far more actionable than either, and it points your content strategy somewhere most local businesses have never looked.

1) Why the headline numbers contradict each other

Three figures circulate, and people cherry-pick whichever supports what they’re selling.

• “AI Overviews appear on 68% of local searches.” True, from Whitespark’s 2026 study of 540 manual queries across six local verticals in three US cities. It’s an average across a query mix that deliberately included informational and hybrid searches.
• “AI Overviews appear on about 40% of local queries.” Also true. Local Falcon’s larger study covered 60,000 queries across 4,423 businesses in 20 countries and logged 40.2%. It ran lower because its methodology included navigational and Boolean queries, where AI Overviews almost never appear.
• “AI Overviews appear on only 7% of local searches.” Also true, if you mean pure, direct local queries. BrightLocal’s tracking puts the overlap between “near me” or city-plus-service queries and AI Overviews at roughly 4% to 8%.

None of these studies is wrong. They measure different mixes of query. The honest range for local business queries overall is roughly 40% to 68% depending on your industry and the kind of queries you count, and quoting any single number without the intent breakdown is how the entire industry has managed to be simultaneously right and useless.

2) The intent split that explains everything

This is the table that should replace every headline statistic you’ve seen. It’s from the Whitespark study, and it’s the most practically useful data in local search right now.

AI Overviews and local packs are inversely related by intent. The more a query looks like “find me a business,” the more likely you get a pack and the less likely you get an AI Overview. The more it looks like “help me understand this before I hire someone,” the more certain the AI Overview becomes.

3) What this means: you’re aiming at the wrong queries

Read the table again and take the strategic consequence seriously, because it inverts the standard advice.

Your Google Business Profile still wins the money queries. “Plumber near me,” “emergency electrician Jaipur,” “dentist open now.” The pack shows for over 90% of those and the AI Overview mostly doesn’t appear. Nothing has been taken from you there. Keep doing the profile work.

But you are almost certainly absent from the queries that come first. Before someone searches “plumber near me,” they searched “why is my water pressure low,” “how much does it cost to replace a boiler in [city],” “how long does a boiler installation take.” Those queries trigger an AI Overview 92% to 97% of the time. And most local business websites contain nothing that could possibly answer them, because most local business websites are entirely “who we are” and “our services.”

That’s the gap. The AI Overview isn’t eating your bottom-of-funnel traffic. It’s owning the top of the funnel that you never contested in the first place, and it’s introducing your competitors to customers a week before those customers ever type a “near me” query.

4) AI local packs, and why they’re worse for you

Alongside AI Overviews, Google has been showing AI-generated local packs, and Sterling Sky’s tracking of them should worry every local business.

Joy Hawkins’ State of Local SEO in 2026 found AI local packs featured 5,943 unique businesses across a query set where regular three-packs featured 18,330. In other words, AI packs surface roughly 32% of the business inventory that the traditional three-pack does. Two thirds of the businesses that would have had a shot at visibility simply don’t appear.

It gets worse on the conversion side:

• AI packs typically show one or two businesses, not three. Fewer slots, same competitors.
• They exclude call buttons entirely, removing the direct-dial path that local businesses have depended on for a decade. A customer who can’t tap to call is a customer who has to take another step, and some won’t.

Meanwhile the paid squeeze is intensifying: Hawkins tracked local pack ad units growing from about 1% of tracked queries in early 2025 to 22% by the end of it, and Local Services Ads from 11% to 31%. The organic local real estate is getting smaller from both directions.

5) What actually gets cited

When an AI Overview answers a local question, it draws from three kinds of source, and you have different amounts of control over each.

• Google Business Profile data. When AI Overviews appear on local-intent queries, they typically cite profile data. Your hours, your category, your services. This is the one you fully control, and it’s covered in Google Business Profile optimization.
• Your own website. Local AI Overviews cite business websites directly, which is the whole reason the content layer in the next section matters.
• Third-party platforms. Reddit, Yelp, review sites, local publications. Overviews routinely scrape and cite these, which means a strong footprint of unstructured citations, mentions of your business on sites that aren’t yours, buys you prominence in AI Overviews that your own site alone cannot. This is exactly the signal local link building builds.

And the prize is real. Seer Interactive’s 2026 analysis found brands cited in an AI Overview earn roughly 120% more organic clicks per impression than uncited brands on the same queries. Being in the box is worth substantially more than merely being on the page beneath it.

The flip side, from the same body of research: organic click-through rate on queries where an AI Overview is present dropped sharply, in one study from 1.76% to 0.61%. If an AI Overview appears and you’re not in it, the page below it is worth much less than it used to be. There is no neutral position here. You are cited or you are compressed.

6) The content layer nobody builds

This is the actual work, and it’s the reason a small local business can win here while national brands can’t be bothered.

Build a page for every question a customer asks before they’re ready to call you. Four families of query, all of which trigger AI Overviews at very high rates:

Where do you get the questions? Your phone. Your inbox. Your reviews. The things customers ask before they book. You already know all of them; you’ve just never written them down. Also mine your Google Business Profile search queries and Search Console. Full method in local keyword research.

7) How to write a page that gets extracted

The craft is the same discipline as winning a featured snippet, because it is, functionally, the same problem: give a machine a clean, liftable answer.

• Answer in the first two sentences. Directly, specifically, with the number if there is one. Then explain. A page that warms up for three paragraphs before answering cannot be extracted.
• One question per page or per section, with the question as the heading, phrased the way a person would say it.
• Be concrete. Ranges, timescales, real figures, what actually happens. Vagueness is uncitable.
• Localize honestly. Costs and timescales differ by area. Saying so, with your area’s numbers, is exactly the specificity that gets you cited on hybrid queries.
• Use clean structure: headers, short paragraphs, lists and tables where they truly help.
• Add an FAQ of the adjacent questions, answered as directly as the main one.
• Keep it current. Prices from 2023 in an AI Overview make you look wrong in public.

The deeper craft of extractable content belongs to writing content that wins featured snippets and AI Overviews. What’s specific to local is the localization: the same question, answered with your area’s real numbers, is a different and more citable page than the generic national version.

8) The profile side still matters

None of this replaces the fundamentals, and it would be a mistake to read the AI Overview data as a reason to neglect them.

The pack still shows for over 90% of simple local queries, which are the ones with the money in them. AI Overviews on local-intent queries cite Business Profile data. And the same underlying signals, reviews, completeness, accuracy, prominence, feed both surfaces.

So: complete profile, narrowest accurate category, accurate hours, steady review flow, consistent listings. It is the same list it has always been, and it now pays into three channels instead of one. Neglecting local SEO in order to chase AI visibility is a false choice; the strategies overlap almost entirely.

9) When the AI Overview gets you wrong

There’s a real hazard here that the optimization conversation tends to skip.

AI Overviews misrepresent local businesses regularly. Wrong hours, wrong services, wrong prices, occasionally wrong businesses. And consumers largely don’t check: GatherUp’s research found 67% of consumers do not rigorously fact-check AI sources before choosing a local business.

So a hallucinated detail in an AI Overview isn’t a technical curiosity. It’s a customer who called someone else because a summary said you closed at five.

Your defences are unglamorous and they’re the same ones as always:

1. Make every source agree. Your website, your schema, your profile, your listings. See NAP consistency.
2. Publish the facts yourself, clearly, so there’s a correct version to extract.
3. Keep everything current, because stale accurate information becomes fresh misinformation.
4. Check what’s being said about you monthly.

10) Measuring this without losing your mind

You can’t rank-track an AI Overview the way you track a blue link. What you can do:

• Track which of your target queries trigger an AI Overview at all. Several rank trackers now flag this. Split your keyword list into “pack queries” and “overview queries” and treat them as different campaigns, because they are.
• Check citation, not position. For your overview queries, are you cited in the box? That’s the metric.
• Watch impressions against clicks in Search Console. Impressions holding while clicks fall on informational queries is the fingerprint of an AI Overview sitting above you.
• Test manually, monthly. Run your cost, process, and duration questions and see who gets cited. It’s twenty minutes and it’s more informative than any dashboard.

Our free Local AI Visibility Grader checks the AI surfaces alongside your profile and citation health.

11) A warning about all of these numbers

Be sceptical of every figure in this article, including ours.

Semrush tracked more than ten million keywords through 2025 and found AI Overview prevalence at 6.49% of queries in January, peaking at 24.61% in July, then falling back to 15.69% by November. That is not a rollout. That is Google experimenting in public, expanding into query types and then retreating from them.

Which means: any single prevalence number has a shelf life measured in months, studies using different keyword samples produce truly different results and are not directly comparable, and anyone quoting a precise figure with confidence is either selling something or hasn’t checked recently.

What is stable, and what you can plan around, is the shape of it: informational and hybrid queries get AI Overviews, transactional local queries get the pack, and the businesses that answer real customer questions in public get cited on both. Build for that, not for a percentage.

The cost page brief: the page your competitors won’t write

Cost and hybrid queries trigger AI Overviews at close to 97%. “Call us for a quote” contains no information and cannot be cited by anything. This is the outline that can be.

H1:  How much does [service] cost in [city]?

FIRST TWO SENTENCES  (this is the bit that gets lifted)
"[Service] in [city] typically costs between [X] and [Y],
 with most jobs landing around [Z]. The three things that
 move the price are [A], [B] and [C]."

THEN:
1. A price table: job type | typical range | what's included
2. What makes it cheaper / more expensive (the honest list)
3. A worked example: a real job, itemised, with the total
4. What's NOT included (callout, parts, VAT, out-of-hours)
5. How [city] differs from the national average, and why
6. How to get an accurate number for your situation
7. FAQ: the six adjacent cost questions people ask
8. Last updated: [date]  <- prices go stale in public

RULES
- Give a real range or don't publish the page.
- Currency, units and dates explicit.
- No "prices vary" as an answer. That is not an answer.

How to find the questions in twenty minutes

• Your phone. What do people ask before they book? Those are the queries. You already know all of them.
• Your quotes and emails. The clarifying questions customers send back.
• Your reviews. What people mention being surprised by, good or bad.
• Google autocomplete. Type “how much does [your service]” and read what appears. Then “how long does”, “do I need”, “should I”.
• People Also Ask. Open one, and it breeds more. Collect twenty.
• Your Google Business Profile search queries. Real demand, from real local searchers, for free.

You are not brainstorming. You are transcribing. The questions already exist; they’ve just never been written down anywhere a machine can read them.

12) Sources used for this guide

FAQ: local AI Overviews

Conclusion: fight where the fight actually is

The local AI Overview story isn’t that Google is taking your “near me” traffic. It mostly isn’t, and the pack is still doing its job. The story is that a whole layer of earlier questions, the ones about cost, process, and timing, is now answered by a summary box, and your competitors are being introduced to your customers inside it while your website talks about your commitment to quality.

Write the pricing page. Write the “how long does it take” page. Answer in the first two sentences, use your area’s real numbers, and keep them true. That’s the entire strategy, it costs nothing but honesty, and almost nobody in your market is doing it.

Then check what the assistants outside Google are saying about you: how to get your local business recommended by ChatGPT.