Insights
AI answers more searches, but local still wins
8 min read

AI and local search are splitting the work. AI answers now handle most "how does it work" and "what does it cost" questions, but searches with local intent, like "plumber near me" or "DUI lawyer," still lead to the map results, calls and websites of nearby businesses. Hiring someone means choosing a licensed, reviewed business you can reach, and the AI tools answering local questions draw on that business's Google Business Profile, reviews and website.
This article covers where AI is taking over, what the data says about local queries, why the split exists, and what to do about it. It also covers what is still unknown.
Where AI answers are taking over
AI Overviews are the summaries Google places above its normal results. Semrush analyzed more than 10 million keywords in 2025 and found they appeared for 6.49% of queries in January, nearly 25% by July, and about 15.69% by November. Early on, roughly 91% of the queries that triggered one were informational. By October that share had dropped to about 57%, as commercial and navigational queries picked up.
Two takeaways. First, AI answers are growing and are spreading beyond pure information questions. Second, there is no public equivalent of this data for ChatGPT, Perplexity or Google's AI Mode, so nobody outside those companies can say exactly how many searches they absorb.
How AI and local search split the work
What the data shows
Whitespark, a local SEO research firm, studied AI Overviews on local business queries. According to its research, as summarized by Search Engine Land, AI Overviews appeared for 68% of local searches while local packs (the three-business map block) appeared for 39%. That sounds like AI is winning local. The breakdown by wording says otherwise:
| Query type | Example | AI Overview | Local pack |
|---|---|---|---|
| Simple local intent | "personal injury lawyers in Phoenix" | 15% | 93% |
| Informational | "how long does an eye exam take near me" | 92% | 6% |
| Hybrid | "average cost of dental implants in Phoenix" | 97% | 17% |
Figures are Whitespark's, as reported by Entrepreneur and Search Engine Land. The 68% average is driven by informational and hybrid queries. The queries a business actually hopes to win, where someone wants a provider, still go to the map.
Synup's study of 53,900 Google results pages across 49 kinds of local business points the same way. It found an AI Overview on 4% of desktop pages for "[category] near me" searches, but on 75% of pages for searches that named a specific need. The three-business map pack appeared on 91% to 97% of pages whatever the wording.
Why the wording changes the answer
Google decides what a person is trying to do. "Why is my water heater making a popping sound" is a request for an explanation, so a written answer fits. "Water heater repair near me" is a request for a vendor, so a list of businesses with phone numbers fits.
Cost questions sit in between. "How much does a Botox appointment cost in Dallas" or "how much does an estate plan cost" mix information with buying intent, which is why they trigger AI answers so often. A business that publishes honest price ranges can be the source those answers draw from.
Why people still choose a nearby business
This part is reasoning rather than measurement, but it explains the numbers. A hire is a commitment with real stakes, and it has four requirements an AI summary cannot fully meet:
- Proximity. A plumber has to drive to the house. A med spa patient has to sit in the chair. Google's own documentation names relevance, distance and prominence as the main local ranking factors. Distance requires a database of places, not a general-purpose text model.
- Licensing. A homeowner wants a licensed electrician. A client wants an attorney admitted in their state or a CPA in good standing. Those checks happen through a business's profile, website and state records.
- Social proof. Reviews carry the trust. Someone choosing a plastic surgeon reads what past patients wrote, not a paragraph generated about "top surgeons."
- A way to act. A burst pipe at 9 p.m. needs a call button. Google's map results provide one. Synup also found that about three in ten local pack listings it observed had no direct website link, so for many buyers the profile is the whole first impression.
Where AI gets its local answers
Different tools lean on different sources, and the evidence quality varies. Much of what follows comes from agency and vendor research, not from the platforms, so treat it as a strong hint rather than a rule.
| Surface | What it appears to lean on | Confidence |
|---|---|---|
| Google map pack | Business Profile, reviews, distance | High (Google documents this) |
| Google AI Overviews | Web pages plus business data | Medium |
| Google AI Mode | Business Profile cards with ratings, hours, a call button and review-based highlights | Medium (vendor analysis) |
| ChatGPT | Web search, licensed listings such as Yelp, business sites | Low to medium |
| Perplexity | Web content; sourcing is less documented | Low |
Reviews show up everywhere. Insites studied 10,000 US local businesses. As reported by Graphem, businesses recommended by both ChatGPT and Perplexity averaged 133 Google reviews, while ignored ones averaged 11. Their profiles were also far more complete (77.8% against 31.5%). That is a correlation. It does not prove 133 reviews is a threshold, and well-run businesses tend to have both reviews and complete profiles.
Whitespark's 2026 Local Search Ranking Factors report surveyed 47 local SEO experts. It is an opinion survey of practitioners, not Google's algorithm. It still found that a dedicated page for each service ranked first for local organic results and second for AI visibility, and that review signals matter for both the map and AI answers.
What to do about it
Protect the map result first
The map pack is where local leads still come from, so it comes first. Choose the most specific primary category, list every service you perform, keep hours and service areas accurate, and add real photos of your work, team and location. Our guides on what local SEO is and how Google picks HVAC map pack results cover the mechanics, and local SEO and Business Profile management is the work itself.
Keep reviews steady, recent and specific
A burst of 50 reviews followed by silence looks different from a steady flow. Ask every satisfied client, make it easy with a direct link, and reply to every review, including the bad ones. A review that mentions the service and the town ("replaced our furnace in Plano") gives both Google and AI tools more to work with. Follow Google's review policies: no incentives and no filtering out unhappy clients. How roofers can get more Google reviews shows what that looks like in practice.
Make your website the source AI quotes
AI tools need something clear to quote. Build one page per service, and one per location where you really serve it. Open each with a plain answer to the question a client would ask, then add specifics: licensing, price ranges, timelines, what the first visit involves, and the areas you serve. Honest service and procedure content is the kind that gets quoted in a hybrid search like "average cost of a repipe in Tucson."
Keep your name, address and phone number identical across your website, Business Profile, Bing Places, Yelp and the directories relevant to your trade. Inconsistent details give AI tools a reason to pick someone else. For the tool-by-tool picture, see AI search optimization and our piece on how home service companies get recommended by ChatGPT.
Measure leads, not rankings
Track calls, form submissions, direction requests and booked consults or jobs, by source. Then test it yourself: search your top five services with and without "near me" in Google, try AI Mode, and ask ChatGPT and Perplexity. If you want an outside read of where you stand across all of them, SEOS offers a free visibility audit.
What is still uncertain
- Google keeps changing the layout. AI Overviews, AI Mode and the map pack are all being tested. A split that holds today may not hold in a year.
- Source weighting is not public. Platforms do not publish how much a Business Profile, a review or a web page counts, so every figure above comes from outside observation.
- Studies are snapshots. They cover specific cities and weeks, and they measure what appears, not what people click or call.
- Local intent may shift. Conversational tools may eventually take on more "find me a provider" requests. The practical response is the same: be the business those tools can verify and recommend.
Frequently asked questions
Are AI answers replacing "near me" searches?
Not so far. Whitespark found AI Overviews on 15% of simple local-intent queries, while the local pack appeared on 93%. AI answers are far more common on informational questions like how long a service takes or what it costs.
Do I still need a Google Business Profile if people use ChatGPT?
Yes. Google's map results and AI Mode work from Business Profile data, and your profile is also a source other tools can read. A complete profile with current reviews is the lowest-cost visibility asset you have.
Which matters more for AI visibility, my website or my reviews?
Both matter, and the research does not give a clean ratio. Reviews and profile data seem to weigh heavily for map results and Google's AI Mode. Clear service pages that answer specific questions look more important for AI Overviews and for tools like ChatGPT and Perplexity.
Should I publish prices on my website?
Publish ranges and explain what moves the price. Cost questions often trigger AI answers, and a page with a clear, honest range gives those answers something to cite. A range with context also pre-qualifies the people who call.
How fast will changes show up?
Business Profile fixes and new reviews can influence map results within weeks. New service pages usually take longer to be crawled, ranked and quoted. Nobody can promise a date, because platforms refresh their data on their own schedules.