Local SEO Still Matters: Where the Map Pack Holds and Which Sources AI Search Cites

New BrightEdge data shows where Google's Local 3-Pack still appears on almost every local-intent keyword we track, which sources ChatGPT and Google AI Mode cite on restaurant and travel prompts, and what OpenAI's new business profiles mean for marketers.

Local SEO remains central to how customers find nearby businesses on Google.

An AI Overview now appears on most of what BrightEdge tracks. Keywords with clear local intent are where that stops. Across three groups cut from the tracked set by wording, an AI Overview appeared on 13.9% of near-me keywords, 13.9% of hotel keywords and 16.6% of local-service keywords. Google's Local 3-Pack appeared on 94.6%, 99.3% and 93.0% of those same three groups.

For a team managing locations, listings and organic visibility, that means Google is still returning local business results when a query signals someone wants a nearby provider, and the inputs to those results are the ones a business controls.

AI search adds a second question. Separate BrightEdge research into restaurant and travel prompts shows ChatGPT citing review and booking platforms, including TripAdvisor, OpenTable and Yelp, more often than the businesses themselves. Those are inputs a business does not control.

The priority is to maintain local SEO while also checking how business information appears across AI search experiences and which sources those systems cite.

These figures describe BrightEdge tracked keywords, United States desktop, from one fixed collection location. It is a keyword set rather than a share of all searches, so it cannot be read as a market rate, and it counts which features appeared rather than clicks, calls or bookings.

Finding 1: which local searches kept the local pack

Three cuts of the tracked set, taken by keyword wording rather than by topic, because wording is what signals local intent.

Keyword cutLocal 3-PackAI Overview
Near me, nearby, closest94.6%13.9%
Hotels in [place]99.3%13.9%
Local-service category words93.0%16.6%

Table 1. Share of keywords in each cut where the feature appeared at collection. Source: BrightEdge tracked keywords, October 2026.

The two columns are counted separately. Each asks whether that feature was on the page, so the two percentages in a row are not parts of one total and a keyword can carry both, one or neither. On the near-me cut, 9.6% carried both.

The pattern holds across three different kinds of local phrasing. A searcher who typed a proximity word, or a place name alongside a business category, got a map and three businesses far more often than a generated summary.

Why this matters now

For most of what a marketing team tracks, an AI Overview is now the default case to plan for. On the keywords with clear local intent it was between 13.9% and 16.6% across the three cuts, so those terms are where planning for a generated answer is not the starting assumption.

That expansion has never been uniform, and BrightEdge has found the same thing in two earlier categories. A three-year study of healthcare found Google widening AI Overview presence across clinical content while pulling back on local and sensitive health queries. Annual snapshots of finance found Google applying comparable category-specific caution to money topics.

What this research adds is that the pattern is not confined to a regulated category but follows the query itself. Across general consumer categories, where the wording signals that somebody wants a nearby business, the pack is what came back.

Whitespark reported AI Overviews on 15% of local-intent queries against 92% of informational ones in May 2025, using its own query set.

What separates a local search from an informational one

On city-modified keywords the result reversed. 84.2% carried an AI Overview with no pack, 14.0% a pack with no AI Overview, and 1.8% neither of the two.

A place name in the keyword does not by itself signal that the searcher wants a nearby business. Google appears to read what the searcher is asking for, and a city name alone does not answer that.

The same business category can sit on both sides. "Dentist near me" behaves like a local search. "How much does a crown cost" behaves like an informational one, and informational local questions are where the AI Overviews in this set are.

"Things to do in Austin" and "best time to visit Rome" carry local words and informational intent, and Google answered them with an AI Overview. This is why published figures on AI Overviews in local search disagree so widely: studies reporting high coverage are usually counting informational queries. The pattern here applies to keywords with clear local intent, not to anything with a place name in it.

Finding 2: which sources ChatGPT and Google AI Mode cited on restaurant and travel prompts

We also examined which sources ChatGPT and Google AI Mode cited for restaurant and travel questions. These are broad category prompt groups taken on prompts rather than keywords, and they carry no local-intent filter, so they are not a matched comparison with Finding 1.

Both engines were read on the same prompt groups in the same week.

Prompt groupEngineMost cited domainsAll other sources
RestaurantsChatGPTTripAdvisor 24.6%, OpenTable 19.6%, Yelp 15.5%40.3%
RestaurantsGoogle AI Modegoogle.com 70.2%29.8%
TravelChatGPTTripAdvisor 45.3%, Expedia 19.9%, Hilton 16.0%18.8%
TravelGoogle AI Modegoogle.com 55.3%, TripAdvisor 11.2%33.5%

Table 2. Share of citations by domain, each row totalling 100%. Named domains are the top cited in that group. All other sources is everything else and is not broken out, so the figure is larger where fewer domains were pulled. Source: BrightEdge AI Brand Presence, October 2026.

The two engines draw on different places. ChatGPT's citations spread across several review and booking platforms. Google AI Mode's concentrate on one domain: on restaurant prompts google.com alone takes 70.2%, more than ChatGPT's top three put together at 59.7%.

For marketers, the findings highlight two areas to manage: local search visibility and the sources influencing AI answers. On Google, businesses can manage important details in their Google Business Profiles, including categories, hours and location information. In AI search, third-party reviews and booking platforms can also contribute information that shapes how a business is represented.

AI Overviews and Google AI Mode are separate surfaces, so a figure from one cannot be quoted as a figure from the other. The presence figures in Finding 1 are AI Overviews. The citation figures here are AI Mode and ChatGPT. AI Catalyst reports which domains each engine cites.

What has changed: ChatGPT business profiles

OpenAI introduced Businesses in ChatGPT on November 10, 2026. Two parts of it matter differently to a brand.

Business Hub is a claimable profile. Eligible businesses can claim and verify their profiles through it, subject to location, business type and verification requirements. Claiming a profile does not guarantee placement, recommendations or sales.

Business Agents are the transactional part, and they are currently available to eligible US ecommerce and online service businesses. OpenAI has said it intends to expand to more categories without naming a date.

Eligibility criteria are set by OpenAI and change, so check the current rules against OpenAI's own documentation before acting on them. The practical read today is that more businesses can claim a profile than can run an agent.

The marketing implication

Local SEO remains the foundation on Google. Managing how a brand appears in AI answers is a second discipline, answer engine optimization or AEO, with different inputs and different sources behind it. Here is how to audit both.

1. Pull your local terms and split them by intent. Separate proximity and category terms ("dentist near me", "hotels in Austin") from informational ones ("how much does a crown cost"). In Data Cube X, check which of each carry a Local 3-Pack and which carry an AI Overview. The two sets behave differently and a blended number hides that.

2. Strengthen your local search presence. For keywords that trigger Google's Local 3-Pack, audit your Google Business Profiles, business categories, location details, customer reviews and relevant location pages. BrightEdge Local SEO helps teams manage listings, maintain consistent business information, monitor Google Business Profile performance and manage reviews across locations.

3. Read the AI answers for your category, then read their sources. Run your top category questions through ChatGPT and Google AI Mode and record which domains are cited. AI Catalyst reports this across prompts rather than one answer at a time.

4. Audit your pages on the platforms that get cited. Where TripAdvisor, Yelp, OpenTable or a booking platform is a cited source for your category, your page on that platform is part of what an engine reads. Check the description, categories, hours and photos the same way you would check a landing page.

5. Claim and verify your ChatGPT Business Hub profile, subject to the eligibility criteria below.

6. Report pack, AI Overview and AI citations as three numbers. Each answers a different question: whether you are in the map, whether an AI Overview appeared, and which sources the AI answer drew on. Optimizing for Google AI Overviews covers the AI side of that reporting.

Methodology

Feature presence comes from BrightEdge tracked keywords, October 2026, United States desktop, collected from one fixed location. Citation shares come from BrightEdge AI Brand Presence for the Restaurants and Travel prompt groups, in the same period.

Presence means the feature appeared on the results page. A citation share is a domain's share of the citations an engine returned for a prompt group. Neither measures clicks, calls, bookings or revenue, and neither shows what a searcher did next.

Two scope notes travel with the figures. The presence figures are AI Overviews, while the citation figures are ChatGPT and Google AI Mode, which are separate surfaces and are labelled separately throughout. The Restaurants and Travel prompt groups include prompts with no local intent, so those shares describe the categories rather than local queries alone, and the two sets of figures are not a like-for-like comparison. Both engines were read on the same prompt groups; prompt-level equivalence within those groups is not established here.

The city-modified shares are reported against every city-modified keyword in the cut. 84.2% carried an AI Overview and no pack, 14.0% a pack and no AI Overview, and 1.8% neither of the two. Nothing is excluded from the base.

Third-party context: Whitespark local study, May 2025, using its own query set. Businesses in ChatGPT facts from OpenAI, November 10, 2026.

Key takeaways

FindingEvidence
Keywords with clear local intent kept the pack93.0% to 99.3% across three cuts of the tracked set
The same keywords carried an AI Overview far less often13.9% to 16.6% across the three cuts
The two features seldom appeared togetherOn city-modified keywords, 84.2% had an AI Overview with no pack against 14.0% the reverse
A place name alone does not signal local intent84.2% of city-modified keywords carried an AI Overview with no pack
ChatGPT returned review platforms on restaurant promptsTripAdvisor 24.6%, OpenTable 19.6%, Yelp 15.5% of Restaurants citations
Google AI Mode mostly cited Google70.2% of Restaurants citations, 55.3% of Travel
The presence path in ChatGPT is wider than the transaction pathHub claiming open to eligible businesses subject to location, type and verification; Business Agents currently US ecommerce and online service

Frequently asked questions

What is the Google Local 3-Pack, and how is it different from an AI Overview?

The Local 3-Pack, also called the map pack, is the block of three local business listings Google shows alongside an interactive map when a query has local intent. It sits near the top of the results page, usually above the standard organic links. It is assembled from Google Business Profile and Maps data rather than from the page content that drives standard organic results.

An AI Overview is a generated summary in roughly the same position, written from indexed sources. On the local-intent keywords in this set the pack appeared on 93% to 99% and an AI Overview on 14% to 17%, so they were usually alternatives rather than companions.

Do AI Overviews show on near-me searches?

Rarely. On the near-me, nearby and closest keywords tracked in October 2026, an AI Overview appeared on 13.9% while the local pack appeared on 94.6%.

Why do "things to do in [city]" searches get an AI Overview when "dentist near me" does not?

Intent. A discovery question is informational and Google answered it with a summary. A proximity search signals that the searcher wants a business, and Google returned the pack. Intent is what separates them, which is why 84.2% of city-modified keywords in this set carried an AI Overview with no pack.

Which sources did ChatGPT cite for restaurant and hotel questions?

Review and booking platforms led. On Restaurants prompts it cited TripAdvisor 24.6%, OpenTable 19.6% and Yelp 15.5%. On Travel prompts, TripAdvisor 45.3%, Expedia 19.9% and Hilton 16.0%. These prompt groups include prompts with no local intent, so the figures describe the categories rather than local queries alone.

Can a local business use ChatGPT Business Agents?

Business Agents are currently available to eligible US ecommerce and online service businesses. Claiming a Business Hub profile is open more widely, to eligible businesses subject to location, business type and verification requirements. Eligibility criteria are set by OpenAI and change, so check OpenAI's documentation for the current position.

Is Google AI Mode the same as an AI Overview?

No, and the difference matters for anyone reading citation data. They are separate surfaces, and in this research AI Mode cited google.com far more often. A figure from one cannot be quoted as a figure from the other.

Does this mean local search is unaffected by AI?

It means keywords with clear local intent are where the Local 3-Pack held most firmly in this tracked set. It does not mean local questions go unanswered by AI, because ChatGPT returns answers for restaurant and travel questions from sources a local brand does not control.

Source: BrightEdge tracked keywords and AI Brand Presence, October 2026, United States, desktop, one fixed collection location. Local study from Whitespark, May 2025. Businesses in ChatGPT facts from OpenAI, November 10, 2026. Part of the Weekly AI & Search Insights series, following last week's edition on Meta crawler traffic.

Download the Full Report

Download PDF

Click the button above to download the full report in PDF format.

Published on November 20, 2026

Subscribe for Weekly Updates