A Sign at Every Exit: Google Maps vs AI Overviews for Local Business Leads

Empty Florida main street at dusk with wet pavement, brick storefronts and one glowing campfire-orange awning
Law Smith, Founder and President of Tocobaga, a Tampa marketing agency

Law Smith

Founder & President, Tocobaga

Law Smith is the Founder and President of Tocobaga, a Tampa-based ROI-focused marketing agency and SMB advisory. He has strategically advised 1,000+ small and medium businesses, executed 600+ integrated marketing campaigns, and written 30+ business plans across industries ranging from professional services to e-commerce to B2B technology. Tocobaga is a Google Partner and Squarespace Gold Partner headquartered in Ybor City.

On September 22, 1978, Arizona opened a new stretch of Interstate 40, and Seligman, a railroad town on Route 66 in northern Arizona, was bypassed. Before that day, roughly 9,000 cars a day rolled down its main drag. Afterward, traffic fell to a trickle, businesses closed and families moved away. Angel Delgadillo, the town barber, later said a person could lie down in the street and not get hit.*

"At first it was so sad, and then I got so angry," Delgadillo told the Associated Press decades later. "Everybody just forgot about us."† Anger is a decent business plan when it comes with a clipboard. In February 1987, his newly formed Historic Route 66 Association of Arizona began lobbying to have the old road declared historic. Most people dismissed the idea. Before the year was out, Arizona did it.* The designation brought highway signs and an annual classic-car Fun Run, which handed interstate travelers the one thing the bypass took away: a reason to exit.†

Down the road, Williams went last, and loudly. On October 13, 1984, a six-mile stretch of I-40 opened around it, making Williams the last Route 66 town in the country to be bypassed.‡,§ About 500 people gathered at the new exit for what was billed as a celebration and functioned as a wake. Bobby Troup, who wrote the song Route 66, arrived by helicopter and sang it with "Williams" added to the lyrics, then later said the day saddened him.‡ Traffic fell and businesses closed. What helped pull cars back was a sign: the Arizona Department of Transportation marked I-40 with "Grand Canyon, next three exits," and the chamber director at the time credited it with drawing motorists off the interstate.‡

Pixar turned the saga into Cars in 2006, with historian Michael Wallis guiding the crew down the old road and voicing the town sheriff.‖ Radiator Springs drew mainly on Seligman, where John Lasseter met the locals, and Delgadillo told Pixar, "We weren't going to take 'no' for an answer."¶ He even turns up after the end credits.*

Every local service business now stands on that main drag. The Google map pack is Main Street: three listings, a pin, a star rating and a phone button, where a homeowner with water coming through the ceiling has always pulled over. AI Overviews, AI Mode and the new Ask Maps are the interstate, built to get the traveler an answer without stopping in town. Nobody in Seligman got a vote on I-40. Nobody on Main Street gets a vote on Gemini.

Law Smith, founder and president of Tocobaga, the ROI-focused lead generation agency and SMB advisory behind SolvingHow.com in Ybor City, Tampa, hears one question from owners nearly every week, in nearly these words: "Are we still getting solid leads from Google Maps and the local pack, or has AI Overview visibility become more important for a service business like ours?" Fair question. Bad frame. Tocobaga's answer is that the road a customer takes depends on what the customer typed. Emergencies still drive down Main Street. Research rides the interstate. Comparison shopping uses both, often in the same afternoon. A business does not get to choose the road. It decides whether there is a sign at every exit.

Are people still finding local businesses through Google Maps in 2026?

Yes, by a wide margin, though Google's share is slipping as new roads open. BrightLocal's 2026 Local Consumer Review Survey of 1,002 U.S. adults found 71 percent used Google to find and evaluate local businesses, down from 83 percent a year earlier, while AI tools such as ChatGPT jumped from 6 percent to 45 percent and Apple Maps climbed from 14 percent to 27 percent.**

Read that as diversification, not collapse. Google Maps serves more than 2 billion people worldwide, according to the Associated Press's coverage of the app's 2026 overhaul.†† On the searches that actually produce service calls, the local pack is close to unavoidable. In a WebFX study that ran 100 home services keywords for HVAC, plumbing, roofing, electrical and garage door companies five times each in May 2026, the local pack appeared on 488 of 500 searches, 97.6 percent, at an average position of 1.5.‡‡ Whitespark's study of 540 local queries across three cities found local packs on 93 percent of local-intent searches.§§

Translate that into trucks. A Brandon pest control company ranking in the pack for "exterminator near me" is not fighting a chatbot for that customer. It is fighting the two other businesses in the pack and whatever ad sits above them. Main Street still has foot traffic. What changed is that it is no longer the only street in town, and the 12-point drop in Google usage is a rough measure of how many travelers now check a second road before they stop.

Two practical implications follow. First, Apple Maps nearly doubled its share in a single year, so a claimed and accurate Apple Business Connect listing has moved from nice-to-have to required for any business whose customers carry iPhones. Second, the 45 percent who used an AI tool did not all quit Google. The two figures add up to well over 100 percent, which means a big share of consumers now check both roads before they pick a business. Tocobaga's earlier explainer, The New Phone Book Is Here, made the case for why the Business Profile matters at all. The 2026 data says it still matters. It is just one sign among several now, and the other signs are getting read.

The road a local customer takes depends on what that customer types, and each road rewards a different kind of sign.
Search intent Example search What shows up first What wins the call
Emergency emergency plumber near me LSAs in position one, then the map pack (WebFX, 2026) Proximity, open hours, recent reviews and a live answer
Routine near me pest control Brandon Map pack on 93% of local-intent queries; AI Overview on 15% (Whitespark) Right primary category, review volume and recency, real photos
Cost research how much does a new roof cost in Tampa AI summary on 60% of question searches (Pew, 2025) Answer-first page with local price ranges the AI can cite
Should-I questions should I get a lawyer after a car accident AI Overview on 97% of hybrid queries (Whitespark) Plain answers, credentials, and named sources on the page
Comparison best estate planning attorney in Clearwater AI summary plus reviews and the pack Specific review text, best-of lists, consistent listings
Conversational Maps termite and rodent company near Seminole Heights, no contract Ask Maps answer inside Google Maps (AP, 2026) Detailed reviews, listed services and complete attributes
Brand check company name plus reviews Business Profile panel and AI review summary Rating, recent replies to reviews, and accurate hours

Do AI Overviews show up for near me searches?

Rarely. In Whitespark's study of 540 local queries across Houston, Phoenix and Denver, AI Overviews appeared on only 15 percent of local-intent searches such as "personal injury lawyers in Phoenix," but on 92 percent of informational searches and 97 percent of hybrid ones such as "should I get a lawyer after a car accident in Phoenix."§§ The 68 percent average that gets quoted on conference stages blends all three, which makes it accurate and nearly useless to a plumber.

Pew Research Center's tracking of real Google searches shows the same pattern from a different angle. Across 68,879 searches by 900 U.S. adults, AI summaries appeared on 18 percent of searches overall, but on 60 percent of searches that used a question word and 53 percent of searches with 10 or more words, versus just 8 percent of one- or two-word searches.‖‖ "Roof leak repair near me" has five words and no question word. "How much does a new roof cost in Tampa" is a nine-word question. One stays on Main Street. The other gets routed onto the interstate.

The honest way to present the numbers is as a range by query type, not as a single figure. For short local searches, AI summaries show up somewhere between 8 percent (Pew's one- and two-word searches) and 15 percent (Whitespark's local-intent queries). For questions, the range runs from 60 percent (Pew's question-word searches) to 92 and 97 percent (Whitespark's informational and hybrid queries). Real estate agents were the outlier in Whitespark's data, with AI Overviews on as many as half of local-intent queries, a reminder that the line moves by industry and that any owner should check a category before assuming.§§

The interstate goes around Main Street, not through it. The driver whose radiator is boiling over still takes the first exit with a gas station. The driver planning a trip three states away reads the guidebook first. A roofer's emergency tarp call and a roofer's $15,000 replacement come from two different travelers, and they are on two different roads.

How much do AI Overviews cut clicks for local businesses?

On the searches where they appear, AI Overviews cut clicks roughly in half. Pew found people clicked a traditional search result on 8 percent of visits when an AI summary appeared, versus 15 percent when none did; they clicked a link inside the summary just 1 percent of the time, and 26 percent of those sessions simply ended, compared with 16 percent without a summary.‖‖

That is on top of a baseline that was already stingy. SparkToro's 2024 study, built on Datos clickstream data from September 2022 to May 2024, found 58.5 percent of U.S. Google searches ended without a click, only 360 of every 1,000 U.S. searches sent a click to the open web, and nearly 30 percent of clicks went to Google-owned properties.¶¶ SparkToro's 2026 follow-up, which switched to Similarweb data, put the zero-click share at about 68 percent of U.S. Google searches for January through April 2026.*** The interstate was already six lanes wide before Gemini repaved it.

Here is what that looks like on a single page. Take a Brandon pest control company with a strong page answering "how much does termite treatment cost." Assume 2,000 people a month see it in the results. Without an AI summary, Pew's 15 percent click rate produces 300 visits. With one, the 8 percent rate produces 160. If 4 percent of visitors call, the page drops from 12 calls to about 6.4. At a 40 percent close rate and a $1,200 average termite job, the business loses about 2.2 jobs and roughly $2,700 a month, or about $32,000 a year, from a page that still ranks first.

Notice where the damage lands. It lands on research pages, the content that answers questions before a customer is ready to call. It does not land on the map pack listing for "exterminator Brandon," where the summary rarely appears. That is Williams after 1984 compressed into one paragraph. The traffic did not stop existing. It stopped stopping.

The fix is not to stop publishing research pages. It is to write them so the AI quotes them, names the business and gives the reader a reason to exit anyway: a local price range the summary cannot fully reproduce, a photo of a real job in Riverview, a calculator, a same-week inspection offer. The click rate drops. The quality of the click that remains goes up.

Which local searches belong to the map pack, Local Services Ads or AI Overviews?

Three kinds of intent, three roads. Emergency and near-me searches such as "emergency plumber near me" belong to the map pack and Local Services Ads; research searches such as "how much does a new roof cost in Tampa" belong to AI Overviews and the pages they cite; and comparison searches such as "best estate planning attorney in Clearwater" are a hybrid where AI summaries, reviews and the pack all get a vote.

Emergency intent is Main Street at rush hour. The customer wants a phone number and a truck, in that order. Speed, proximity, open hours and star rating decide it, and the sale happens on a phone call within minutes. Whitespark's data shows local-intent queries are the most likely to produce a local pack and the least likely to produce an AI Overview.§§ For a roofer after a summer squall, a pest company in termite swarm season or an HVAC contractor in a Tampa August, this lane carries the volume.

Research intent is the interstate. "How much does a new roof cost in Tampa." "Does homeowners insurance cover roof replacement in Florida." "How long does tenting for termites take." These searches trigger AI summaries most of the time, and the customer may read an answer, close the tab and call a company by name a week later. That call shows up in reports as a branded search or a direct call, and the research page that earned it gets no credit. Lower volume, higher ticket, longer fuse.

Comparison intent is the exit ramp itself. A two-partner estate planning firm in Clearwater does not get 2 a.m. emergency calls; nobody needs a revocable trust before sunrise. Its clients type "best estate planning attorney in Clearwater," read whatever summary appears, scan the reviews and then call one or two firms. BrightLocal found 82 percent of consumers read AI-generated review summaries and 42 percent trust AI recommendations as much as traditional reviews.** In this lane, what clients wrote about the firm is the sign, and the AI is reading it aloud to strangers.

The same split shows up inside larger operators. A PE-backed HVAC platform on its second add-on has three lanes at every rooftop: "AC not cooling" calls in July on Main Street, "heat pump vs AC cost Florida" research on the interstate, and "best HVAC company in Brandon" comparisons on the ramp. A platform that runs one playbook for all three lanes is putting up a billboard where it needed a gas station.

Pest control technician's hand reaching for a ringing phone in a truck cab at dawn, tracking local service calls

Where do service business calls actually come from now?

Mostly from Google, split among paid search, the Business Profile and organic results, with AI tools sending a small but unusually qualified trickle. CallRail's analysis of 1.1 million de-identified conversations found Google Ads drove 37 percent of calls and texts, the Google Business Profile 23 percent and organic search 22 percent, and the Business Profile produced 32 percent of sales opportunities, close behind Google Ads at 36 percent.†††

Sit with that last number. A free listing produced nearly as many sales opportunities as the paid channel most owners budget for first. Joy Hawkins of Sterling Sky, who has tracked Business Profile leads across client accounts for years, reports that for most industries the profile drives substantially more leads for small businesses than any other source.‡‡‡ Main Street is not dying. It is underpriced.

Invoca's 2026 home services benchmarks, drawn from a dataset of more than 70 million calls, add the uncomfortable part. Only 52 percent of all inbound calls were answered, 38 percent of answered calls were genuine leads, 45 percent of those leads converted on the call, and 55 percent of businesses never asked the caller to buy or book.§§§ Calls referred by ChatGPT were the most qualified of any channel, with 45 percent of answered calls turning out to be leads and 41 percent of those leads converting, but generative AI still produced very little measurable call volume, and Invoca's own advice was to monitor it rather than move significant budget.§§§

Run that through one month. A home services company that gets 200 inbound calls answers about 104 of them at the benchmark rate. Roughly 40 of those are real leads, and about 18 book. Invoca calculates that five-point improvements in answer rate, lead rate and conversion rate together produce roughly 38 percent more conversions from the same call volume.§§§ For most owners losing sleep over AI Overviews, the bigger leak is a phone ringing on a desk nobody is sitting at.

The ChatGPT traffic looks a lot like the Route 66 Fun Run: fewer cars, better buyers, and every one of them chose to take the exit. That traffic deserves a sign and a clean way to measure it. It does not yet deserve the whole marketing budget, and anyone who says otherwise should be asked to show the call logs.

Why don't Google Business Profile call numbers match call tracking?

Because they count different things. The Business Profile performance report counts calls placed by tapping the profile on a mobile device, so a desktop searcher who reads the number and dials by hand is invisible and repeat taps by the same person are not counted as unique; call tracking counts actual calls from unique callers on any device.‖‖‖ In one Sterling Sky comparison, call tracking logged 380 unique calls in a month while the profile reported 282, about 74 percent, though the month-to-month trend in both matched.‡‡‡

BrightLocal's walkthrough of call tracking for local businesses describes a similar month in which tracking captured 54 calls the profile never reported.‖‖‖ Neither number is a lie. One is a speedometer and the other is an odometer, and owners keep reading the speedometer to decide whether the trip was worth taking.

This matters for the Maps versus AI debate because the most common evidence offered that Maps is dying is a falling call count in the profile dashboard. Before any budget moves, an owner should ask three questions. Did profile views fall too, or only calls? Did tracked calls to the profile's number fall by the same amount? Did the business change its hours, categories or phone routing in that window? Often the decline turns out to be a measurement story wearing a strategy costume.

The fix is mechanical. Put a dedicated call tracking number in the profile's primary phone field, keep the main business line as an additional number so name, address and phone stay consistent with the website, and tag the profile's website link with UTM parameters so form fills from Maps stop hiding inside organic traffic. Then add an "AI assistant" choice to the intake question about how the caller found the business. It is imperfect and self-reported, and it is still better than a guess.

Williams could count the 500 people standing at the exit ceremony. The number that mattered was the cars that took the exit every day afterward, and nobody was counting those. Most service businesses are in the same spot: precise about the vanity number, blind on the one that pays the mortgage.

How does a business rank in the Google map pack in 2026?

Google says local results are based mainly on relevance, distance and prominence, and Whitespark's 2026 Local Search Ranking Factors survey of 47 experts scoring 187 factors put the primary Business Profile category first and the searcher's proximity to the business a hair behind, 227 points to 225.¶¶¶,**** Keywords in the business title came third at 223, and a physical address in the city of search came fourth at 213.****

Read the top four as a map of what an owner can and cannot change. Category is fully controllable, and a mistake Tocobaga sees constantly in audits is a primary category chosen for vanity rather than for the service that makes the money. Proximity is not controllable at all; the shop is where the shop is. Business name keywords are controllable only when the keyword is genuinely part of the real business name, because stuffing extra words into the listing breaks Google's guidelines and invites a suspension. Address in the city of search is a structural disadvantage for service-area businesses that hide their address, which is why a Tampa roofer with a Brandon warehouse can rank better in Brandon than in South Tampa no matter how many blog posts it publishes.

Prominence is where effort pays. Google says more reviews and positive ratings can help a business's local ranking.¶¶¶ Customers have raised the bar at the same time. BrightLocal found 68 percent of consumers will only use a business with at least four stars, up from 55 percent; 74 percent care only about reviews from the last three months; and 47 percent will not use a business with fewer than 20 reviews.** A pest control company sitting at 4.3 stars on 19 reviews, the newest from March, is a diner on Main Street with half its neon burned out. Technically open. Functionally invisible.

Tocobaga's setup checklist, Worth a Detour, covers the profile fields line by line. The ranking reality in 2026 is simpler than any checklist: pick the right primary category, earn a steady stream of recent and specific reviews, keep the hours honest, answer every review, and stop trying to move the building.

Are Local Services Ads worth it compared to the map pack?

For emergency and high-intent home services searches, usually yes, because Local Services Ads sit above everything else when they show: in WebFX's May 2026 study, LSAs appeared on 28.4 percent of 500 home services searches and 49 percent of city-modified ones, and they held position one every time they appeared.‡‡ They do not replace the map pack, which showed up on 97.6 percent of the same searches. They sit on top of it.

On a phone, that matters more than the percentages suggest. An LSA unit, an AI Overview and a row of search ads can stack above the three-pack and push it below the first screen, and any owner can check this in ten seconds by searching the company's top service and city on a phone. When the pack slides down, the business that bought the billboard at the exit ramp gets the first look.

Whitespark's 2026 survey ranked the top LSA factors as budget and bidding, reviews, service selection and response times.**** Three of those four are operational, not creative. LSAs charge per lead rather than per click, which makes the math clean. Suppose a Brandon pest control company pays $45 per LSA lead and closes 35 percent of them. Its acquisition cost is about $129 per customer. If a typical new customer is worth $600 in the first year, every $1 spent returns about $4.67 in first-year revenue before labor and materials. If the company answers slowly and closes only 20 percent, the cost per customer jumps to $225 and the return drops to about $2.67. Same ad. Same budget. The phone decided the ROI.

The honest verdict: LSAs are worth testing for any home service business with strong reviews and someone who answers the phone within a few rings. They make less sense for firms whose clients research for weeks before calling, and they are no substitute for the pack. Williams needed the state's sign on the interstate and storefronts worth stopping at. A paid sign pointing to a closed diner is just an expensive way to annoy people.

Orange bucket catching a ceiling leak on a terrazzo floor, the emergency search behind local map pack calls

What is Ask Maps and how does it pick businesses?

Ask Maps is a Gemini-powered conversational search built into Google Maps, announced March 12, 2026, rolled out first on iPhone and Android in the U.S. and India, and expanded to more than 150 countries and territories in English on August 6, 2026;†††† it answers complex, specific questions by drawing on a database of more than 300 million places and reviews from more than 500 million contributors.†† Google executives declined to say whether ads will eventually appear in its recommendations.††

Ask Maps is the strangest development of the year because it builds the interstate on top of Main Street. The same app that shows the three-pack now lets a customer ask for something like a pest company near Seminole Heights that handles both termites and rodents and does not require an annual contract. No category field holds that answer. Reviews do. A business whose customers wrote that a technician showed up the same afternoon for rats in the attic and never pushed a contract just had its sign painted by its own customers.

That makes review content, not just review count, a ranking asset. BrightLocal found 82 percent of consumers read AI-generated review summaries and 23 percent would rely on them alone.** The practical shift is in the ask. Instead of a generic request for stars, the follow-up message invites the customer to mention what the job was, the neighborhood and what stood out. Scripting the words reads fake and risks running afoul of Google's review policies. Prompting for specifics is just good manners with a long tail.

The same logic covers everything else Ask Maps can read: services listed in the profile, attributes such as weekend hours or financing, photos of real jobs, and the business description. Every field left blank is a question the AI cannot answer on the business's behalf, which means a competitor answers it instead.

Delgadillo's association lobbied for highway signs because travelers needed to be told the old road was worth the detour. In 2026 the highway reads the signs for itself, so the paint has to say something specific.

Did the March and May 2026 core updates change local rankings?

They changed websites more than map listings. Google ran two broad core updates in the first half of 2026, one from March 27 to April 8 and a second from May 21 to June 2 that Search Engine Roundtable's Barry Schwartz judged the larger of the two.‡‡‡‡ A third core update began November 10, with a rollout of up to two weeks.§§§§ Core updates re-rank web pages, so the direct effects land on organic results and on the pages AI Overviews choose to cite, while the map pack leans heavily on Business Profile signals such as category, distance and reviews.

That distinction saves money. When a roofer's map pack position slides during an update window, the first suspects are a competitor's new reviews, a category change, a moved pin or an edit someone suggested to the listing. When the website's research pages slide, the update is a legitimate suspect, and the response is better answers on fewer, stronger pages, not a flood of thin new ones.

The cleanest test takes an afternoon. Compare the 28 days before an update with the 28 days after its rollout ended. Separate branded searches from non-branded ones. Separate profile calls from website calls using tracking numbers. Look at booked jobs, not impressions. A business that skips this step tends to fix the wrong road. Repaving the interstate does not move Main Street, and a pothole on Main Street is not the highway department's fault.

For a management consultancy or a Tampa 3PL that sells to other businesses, the core updates matter more than the pack ever will, because their buyers research on the interstate almost exclusively. For a drain cleaning company, a core update is mostly weather news: worth watching, rarely worth panicking over.

What is the strongest case for ignoring AI Overviews entirely?

The strongest case is that the money is on Main Street and the data says so. AI Overviews appeared on only 15 percent of local-intent searches in Whitespark's study while local packs appeared on 93 percent, and Invoca's 2026 benchmarks found generative AI still drives very little measurable call volume for home services.§§,§§§

Take that argument seriously, because a lot of profitable plumbers make it. Emergency revenue comes from the pack, LSAs and a phone that gets answered. Every hour spent writing a 2,000-word page about attic ventilation is an hour not spent asking for reviews or training the dispatcher. AI answers are hard to track, and Whitespark warned that AI Overviews have a habit of containing misinformation, which makes them a business risk, not just a channel.§§ For a two-truck drain company, ignoring the interstate and polishing the storefront is a defensible strategy.

The case breaks in three places. First, ticket size. Research queries feed the expensive jobs: the roof replacement, the full-system HVAC swap, the estate plan, the commercial pest contract. Those buyers ask questions first, and questions trigger AI summaries 60 to 97 percent of the time across the Pew and Whitespark data.‖‖,§§ Second, the trend line. The share of consumers using AI tools to find and evaluate local businesses went from 6 percent to 45 percent in a year.** Third, the road itself. Ask Maps puts conversational AI inside the Maps app, so ignoring AI now means ignoring the newest lane on Main Street.††

So the right version of the argument is narrower: ignore AI Overviews for emergency keywords, never for the business. The Route 66 towns that pretended the interstate was a fad did not keep their traffic. They kept their pride and lost the motel.

How should a service business split effort between Google Maps and AI search?

Split effort by where revenue comes from, not where leads come from. Tocobaga's starting point for most local service businesses is about 55 percent of marketing effort on Main Street (the Business Profile, reviews, LSAs and answering the phone), 30 percent on the interstate exits (research pages, citations and AI visibility) and 15 percent on measurement, and then the numbers move it.

Here is the arithmetic for a hypothetical Tampa roofing company doing about $4 million a year. Its monthly leads, close rates and average tickets by source look like this:

  • Map pack and Business Profile calls: 80 leads, 30 percent close rate, 24 jobs, $3,500 average ticket, $84,000

  • Local Services Ads: 40 leads, 25 percent close rate, 10 jobs, $3,000 average ticket, $30,000

  • Organic research pages on cost, insurance and repair versus replace: 50 leads, 12 percent close rate, 6 jobs, $15,000 average ticket (mostly replacements), $90,000

  • Google Ads search: 30 leads, 20 percent close rate, 6 jobs, $6,000 average ticket, $36,000

  • Referrals and repeat customers: 20 leads, 50 percent close rate, 10 jobs, $9,000 average ticket, $90,000

That is 220 leads, 56 jobs and $330,000 a month, or about $3.96 million a year. The map pack produces 36 percent of the leads but only about 25 percent of the revenue. The research pages produce 23 percent of the leads and 27 percent of the revenue, because the people who research are the people buying whole roofs.

Now let AI Overviews take 20 percent of the research clicks. Research leads fall from 50 to 40. At a 12 percent close rate, jobs fall from 6 to 4.8. Losing 1.2 replacements at $15,000 each is $18,000 a month, or $216,000 a year, about 5.5 percent of revenue. If the full Pew ratio of 8 clicks for every 15 held across every research query, leads would fall to about 27 and the loss would be roughly $42,000 a month.

Now run the Main Street risk. Suppose a competitor's review push knocks the roofer from first to third in the pack, and pack calls fall 40 percent. Leads drop from 80 to 48, jobs from 24 to 14.4, and revenue by $33,600 a month, or about $403,000 a year. Main Street carries the bigger single risk. The interstate carries the bigger tickets. And the cheapest lever of all sits on the dispatcher's desk: if the roofer captured even a third of the 38 percent lift Invoca attributes to five-point gains in answer, lead and close rates, the map pack line alone would add about three jobs and roughly $10,600 a month.§§§

That is why the split is 55, 30 and 15 rather than all-in on either road. For the two-partner estate planning firm in Clearwater, where comparison and research intent dominate, the split flips to roughly 40 percent Main Street, 45 percent interstate and 15 percent measurement. For the PE-backed HVAC platform with six rooftops, it tilts the other way, about 60, 25 and 15, with a big share of the measurement work spent consolidating call tracking across acquired brands so the board deck stops arguing with the phone bill. The percentages are a starting point. The call log is the referee.

Vintage motel front desk with a pegboard of room keys, one orange key fob, a rotary phone and a pink flamingo

What should a service business do this quarter to win Maps and AI search?

Start with measurement, then protect Main Street, then build the exit signs. In order:

  • Put a call tracking number in the Business Profile's primary phone field, keep the main line as an additional number, and add UTM tags to the profile's website link.

  • Pull the last 90 days of calls and sort them by intent: emergency, research and comparison. That mix decides the split of effort.

  • Fix the phone before the marketing: measure the answer rate, route after-hours calls to a live person or a callback within 15 minutes, and ask every qualified caller to book.

  • Audit the primary category and listed services against the services that make the most money, and change one thing at a time.

  • Send a review request after every completed job that invites customers to mention the service, the neighborhood and what stood out.

  • Publish answer-first research pages with real local price ranges, timelines and insurance details for the three most expensive services.

  • Claim and clean up Apple Business Connect and Bing Places, and pursue mentions on local "best of" lists and industry directories that AI answers cite.

  • Test Local Services Ads on emergency categories with a capped budget, and dispute every lead that is not a real job inquiry.

  • Once a month, search the top 10 money keywords on a phone and screenshot what sits above the pack.

  • After any core update, wait for the rollout to finish, compare 28 days before and after by booked jobs, and only then change anything.

Tocobaga's earlier field guide to AI citations, Phone a Friend, covers the content side in detail: entity consistency, answer-first structure and how to earn a mention from ChatGPT and AI Overviews. The list above is the local layer on top of it. Every item is a sign at an exit. None of them requires guessing which road the customer will take, which is the point. The owner who builds for both roads never has to win the argument about which one matters more.

Should a local business hire help for Google Maps and AI search?

Only if the help reports in calls and booked jobs rather than rankings and impressions, because a map pack position that does not ring the phone is a trophy, not an asset. Tocobaga has advised more than 1,000 small and mid-sized businesses, executed more than 600 integrated campaigns and produced more than 300 websites, and the first thing it builds for a service business is the call tracking that makes the Maps versus AI question answerable with data instead of vibes.

For owners who want a second set of eyes on where their calls actually come from, Tocobaga offers an intro call at (813) 534-1261 or through hello@solvinghow.com at SolvingHow.com. The model is deliberately boring: flat fees, never a percentage of ad spend, no auto-renewing contracts, and the client owns every account, profile, tracking number and asset. There is no Tocobaga name in the footer of a client's website. And Tocobaga is not for everyone; when it is not the right fit, it says so and points the owner toward someone who is.

Will AI Overviews replace the Google map pack?

Not for emergencies, and not soon. The more likely future is that the two roads merge, with Ask Maps running AI through the map itself and AI Overviews borrowing from the same reviews and listings the pack already uses, so the businesses that win will be the ones visible on both.

Williams survived. The town rebounded in the 1990s, when the Grand Canyon Railway was bought and revived and Route 66 nostalgia turned into a tourism economy.‡ Seligman was reborn on the same nostalgia, its businesses built around the very road that had been taken away from it.† Neither town got its traffic back by fighting Interstate 40. They put signs at the exits, gave travelers a reason to stop and made the stop worth it once they arrived.

That is the whole job for a local service business in late 2026. The map pack is still Main Street, and for the roof leak at midnight it is still the only street that matters. AI Overviews and Ask Maps are the interstate, and they already carry the customers with the biggest checks. The owner who asks which road matters more is asking the question Seligman asked in 1978. The better question is the one Angel Delgadillo answered in 1987.

He could not stop the interstate. Nobody could. He made sure that when a traveler on it wondered whether anything between Kingman and Flagstaff was worth the detour, the answer had his town's name in it. The customer picks the road. The business paints the sign.

Is Google Maps or AI Overviews more important for a plumber or HVAC company?

For emergency work, Google Maps and the local pack matter more. Whitespark's study of 540 local queries found AI Overviews on only 15 percent of local-intent searches but local packs on 93 percent. AI Overviews matter for research searches tied to bigger jobs, such as full system replacements, where they appeared on 92 percent of informational queries. Most plumbing and HVAC companies should weight effort toward the map pack while publishing a handful of strong cost and comparison pages.

Do AI Overviews and the local pack appear on the same search?

Sometimes. Whitespark's 2025 case study found local packs on 17 percent of hybrid queries, where AI Overviews appeared 97 percent of the time, and found no reliable correlation between how prominent the pack was and whether an AI Overview appeared. On a phone, an AI Overview, Local Services Ads and search ads can stack above the pack and push it down the screen. Checking the top keywords on a phone each month shows exactly what customers see.

What is the number one ranking factor for the Google map pack?

Whitespark's 2026 Local Search Ranking Factors survey, in which 47 experts scored 187 factors, ranked the primary Google Business Profile category first with 227 points and the searcher's proximity to the business second with 225. Keywords in the business title and a physical address in the city of search followed. Google's own documentation says local results are based mainly on relevance, distance and prominence, with reviews contributing to prominence.

Can ChatGPT send phone calls to a local business?

Yes, though the volume is still small. Invoca's 2026 home services benchmarks found ChatGPT-referred calls were the most qualified of any channel, with 45 percent of answered calls turning out to be real leads and 41 percent of those converting. Invoca also found generative AI produces very little measurable call volume so far and recommends monitoring it rather than moving significant budget. Adding an AI assistant option to intake questions helps track it.

How can a business track calls that come from Google Maps?

The standard method is a dedicated call tracking number placed in the Google Business Profile's primary phone field, with the main business line kept as an additional number so listings stay consistent. Adding UTM tags to the profile's website link separates Maps traffic from organic search. In one Sterling Sky comparison, the profile's own reporting showed about 74 percent of the unique calls that call tracking recorded, so tracked numbers are the more reliable count.

What percentage of Google searches end without a click?

SparkToro's 2026 study, based on Similarweb data for January through April 2026, found about 68 percent of U.S. Google searches ended without a click. Its 2024 study, based on Datos clickstream data, found 58.5 percent, and only 360 of every 1,000 U.S. searches sent a click to the open web. Pew Research Center's 2025 study found that when an AI summary appeared, people clicked a traditional result on just 8 percent of visits, compared with 15 percent when no summary appeared.

Is Ask Maps available to every business?

Ask Maps is a feature for searchers, not a product a business signs up for. Google announced it on March 12, 2026, for the Google Maps mobile app on iPhone and Android in the U.S. and India, and on August 6, 2026 expanded it to more than 150 countries and territories in English. It draws on more than 300 million places and reviews from more than 500 million contributors, so a complete Business Profile and detailed, recent reviews are how a business shows up in its answers.

Should a small business pay for Local Services Ads if it already ranks in the map pack?

Often yes for home services, because LSAs sit above the pack. WebFX's May 2026 study found LSAs on 28.4 percent of home services searches and 49 percent of city-modified ones, always in the top position. A business already in the pack can use LSAs to own more of the first screen on mobile. The test is cost per booked job: if LSA leads close at a healthy rate and the phone gets answered fast, the ads usually pay.

Sources
*
Las Vegas Review-Journal. "Arizona man, 91, still relishes role as Angel of Route 66." Las Vegas Review-Journal, July 1, 2018. Reports that Seligman was bypassed by Interstate 40 on September 22, 1978, that daily traffic fell from about 9,000 cars to a trickle, that Angel Delgadillo's Historic Route 66 Association of Arizona began lobbying in February 1987 and Arizona listed Route 66 as historic in November 1987, and that Delgadillo appears after the end credits of Pixar's Cars. https://reviewjournal.com/local/local-nevada/arizona-man-91-still-relishes-role-as-angel-of-route-66
†
Arrillaga, Pauline. "At 84, Guardian Angel Still Watching Over Route 66." Associated Press via Daily Herald, May 29, 2011. Quotes Delgadillo ("Golly Moses. At first it was so sad, and then I got so angry" and "Everybody just forgot about us") and reports that the association's lobbying led to Historic Route 66 highway signs and an annual classic-car Fun Run, and that Seligman was reborn through Route 66 tourism. https://www.dailyherald.com/20110529/news/at-84-guardian-angel-still-watching-over-route-66/
‡
Warnick, Ron. "Route 66 in Williams was bypassed 40 years ago." Route 66 News, December 16, 2024. Reports the October 13, 1984 six-mile I-40 bypass of Williams, about 500 people at the Grand Canyon Boulevard exit, Bobby Troup's helicopter arrival and his later regret, the decline after the 1985 decommissioning, the 1990s rebound with the revived Grand Canyon Railway, and ADOT's "Grand Canyon, next three exits" sign. https://www.route66news.com/2024/12/16/route-66-in-williams-was-bypassed-40-years-ago/
§
National Park Service. "Route 66: Arizona." U.S. Department of the Interior, December 3, 2021. States that the last segment of Route 66 in the nation was bypassed by the interstate at Williams, Arizona, in 1984, and describes the grassroots preservation effort that followed. https://www.nps.gov/articles/000/route-66-arizona.htm
‖
Warnick, Ron. "New York Times tells about creation of 'Cars.'" Route 66 News, May 20, 2006. Summarizes Phil Patton's New York Times report that historian Michael Wallis led the Pixar team on two Route 66 tours and voiced the sheriff of Radiator Springs, a 1949 Mercury. https://www.route66news.com/?p=924
Finnigan, Reese. "Seligman, AZ: The Real-Life Radiator Springs." Arizona Preservation Foundation on Google Arts and Culture, accessed October 26, 2026. Identifies Seligman as the main basis for Radiator Springs, notes that John Lasseter visited Seligman for research, and quotes Delgadillo's 2006 remark to Pixar, "We weren't going to take 'no' for an answer." https://artsandculture.google.com/story/seligman-az-the-real-life-radiator-springs-arizona-preservation/RAUhn8A2gZghmA?hl=en
**
BrightLocal. "Local Consumer Review Survey 2026." BrightLocal, February 11, 2026. Survey of 1,002 U.S. adults: 71 percent used Google to find and evaluate local businesses (down from 83), 45 percent used AI tools (up from 6), 27 percent Apple Maps (up from 14); 42 percent trust AI recommendations as much as reviews; 82 percent read AI review summaries and 23 percent would rely on them alone; 74 percent care only about reviews from the last three months; 68 percent require four stars; 47 percent avoid businesses with fewer than 20 reviews. https://www.brightlocal.com/research/local-consumer-review-survey/
††
Liedtke, Michael. "Google overhauls its Maps app, adding in more AI features to help people get around." Associated Press via ABC13, March 12, 2026. Reports Ask Maps rolling out on the Google Maps mobile app in the U.S. and India, drawing on more than 300 million places and reviews from more than 500 million contributors, Google Maps' more than 2 billion users, and executives declining to say whether ads will appear in its recommendations. https://abc13.com/post/google-overhauls-maps-app-adding-more-ai-features-help-people-get-around/18706200/
‡‡
Velasquez, Albert Dandy, and Ryan Maake. "What Dominates Home Services Search: 2026 Study of Local Services Ads, Google Ads and the Local Pack." WebFX, July 8, 2026. Across 500 home services searches in May 2026, LSAs appeared on 28.4 percent overall and 49 percent of city-modified searches, always in position one; the local pack appeared on 97.6 percent at an average position of 1.5. https://www.webfx.com/blog/home-services/local-services-ads-vs-google-ads-vs-local-pack-study/
§§
Ellis, Miriam. "Case Study: The Prevalence of AI Overviews in Local Search." Whitespark, May 12, 2025. Across 540 queries in Houston, Phoenix and Denver, AI Overviews averaged 68 percent but appeared on 15 percent of local-intent, 92 percent of informational and 97 percent of hybrid queries; local packs appeared on 93 percent of local-intent queries; real estate agents reached 50 percent; the study warns AI Overviews can contain misinformation. https://whitespark.ca/blog/case-study-the-prevalence-of-ai-overviews-in-local-search/
‖‖
Chapekis, Athena, and Anna Lieb. "Google users are less likely to click on links when an AI summary appears in the results." Pew Research Center, July 22, 2025. Across 68,879 searches by 900 U.S. adults, 18 percent showed an AI summary; traditional-result clicks fell from 15 percent to 8 percent of visits, summary links drew 1 percent, and 26 percent of sessions ended versus 16 percent; summaries appeared on 60 percent of question-word searches, 53 percent of 10-plus-word searches and 8 percent of one- or two-word searches. https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
¶¶
Goodwin, Danny. "Nearly 60% of Google searches end without a click in 2024." Search Engine Land, July 2, 2024. Reports SparkToro's study of Datos data from September 2022 to May 2024: 58.5 percent of U.S. Google searches ended without a click, 360 of every 1,000 U.S. searches sent a click to the open web, and nearly 30 percent of clicks went to Google-owned properties. https://searchengineland.com/google-search-zero-click-study-2024-443869
***
Fishkin, Rand. "In 2026, Less than One Third of Google Searches Still Send a Click." SparkToro, June 8, 2026. Based on Similarweb data for January through April 2026, 68.01 percent of U.S. Google searches ended without a click. https://sparktoro.com/blog/in-2026-less-than-one-third-of-google-searches-still-send-a-click/
†††
CallRail. "CallRail Releases Report Benchmarking Marketing Efforts for Small Businesses." Business Wire, January 14, 2025. Across 1.1 million de-identified conversations, Google Ads drove 37 percent of calls and texts, Google Business Profile 23 percent and organic search 22 percent; the Business Profile produced 32 percent of sales opportunities versus 36 percent for Google Ads. https://www.businesswire.com/news/home/20250114793850/en/CallRail-Releases-Report-Benchmarking-Marketing-Efforts-for-Small-Businesses
‡‡‡
Hawkins, Joy. "How to Increase Leads From GMB and Get More Customers." Sterling Sky, September 22, 2026. Reports that for most industries the Business Profile drives substantially more small business leads than any other source, and that in one comparison call tracking recorded 380 unique calls while profile insights showed 282, with matching month-to-month trends. https://www.sterlingsky.ca/leads-google-my-business/
§§§
Invoca. "2026 Home Services Benchmarks." Invoca, July 2026. From a dataset of more than 70 million calls: 52 percent of inbound home services calls answered, 38 percent of answered calls were leads, 45 percent of leads converted, 55 percent of businesses did not ask callers to book; ChatGPT-referred calls had a 45 percent lead rate and 41 percent conversion; generative AI call volume remains small; five-point gains in three rates yield about 38 percent more conversions. https://invoca.com/reports/the-invoca-home-services-lead-conversion-benchmarks-report-2026
‖‖‖
Nielson, Colan. "Call Tracking for Local Businesses: How to Set It Up and Why You Should." BrightLocal, December 14, 2021, updated October 31, 2025. Explains that Business Profile insights count only calls placed from mobile devices and do not count unique callers, and describes a month in which call tracking captured 54 calls the profile did not report. https://www.brightlocal.com/learn/call-tracking-for-local-businesses/
¶¶¶
Google. "How to improve your local ranking on Google." Google Business Profile Help, accessed October 26, 2026. States that local results are based mainly on relevance, distance and prominence, and that more reviews and positive ratings can help a business's local ranking. https://support.google.com/business/answer/7091?hl=en
****
Shaw, Darren. "2026 Local Search Ranking Factors." Whitespark, November 6, 2025. Survey of 47 experts scoring 187 factors: primary Business Profile category (227), proximity to the searcher (225), keywords in the business title (223) and physical address in the city of search (213) lead the local pack; top Local Services Ads factors are budget and bidding, reviews, service selection and response times. https://whitespark.ca/local-search-ranking-factors/
††††
Daniel, Miriam. "Ask Maps gets more helpful with food ordering and more." Google, The Keyword, August 6, 2026. Reports Ask Maps rolling out in Australia, Brazil, Canada, Indonesia, Japan and Mexico plus more than 150 countries and territories in English, with food ordering, hotel and event discovery launching first in the U.S. https://blog.google/products-and-platforms/products/maps/order-food-in-ask-maps/
‡‡‡‡
Schwartz, Barry. "Google May 2026 Core Update Has Completed Rolling Out." Search Engine Roundtable, June 2, 2026. Reports the March 2026 core update ran from March 27 to April 8 and the May 2026 core update from May 21 to June 2, and that the May update felt larger. https://www.seroundtable.com/google-may-2026-core-update-done-41435.html
§§§§
Google Search Status Dashboard. "November 2026 core update." Google Search Central, November 10, 2026. States that Google released the update on November 10, 2026, that it applies globally and to all languages, and that the rollout may take up to two weeks. https://status.search.google.com/incidents/WLxcJyC6jMWMfryxsKMc
Next
Next

The Zagat Machine: What Gets a Local Business Recommended by ChatGPT, Gemini and AI Overviews