---
title: "The Local Answer Gap: what 7,583 Google AI answers reveal about AEO for local businesses"
author: Local AI Registry Research Team
publisher: Local AI Registry (https://localairegistry.com)
date: 2026-11-07
canonical: https://localairegistry.com/research/local-answer-gap
license: CC BY 4.0
---

[Local AI Registry](https://localairegistry.com) › [Research](https://localairegistry.com/research) › The Local Answer Gap

Local AI Registry research · Published November 7, 2026

# The Local Answer Gap: what 7,583 Google AI answers reveal about AEO for local businesses

By the [Local AI Registry Research Team](https://localairegistry.com/research/local-answer-gap#authors)

The short answer

Local AI Registry’s analysis of 7,583 Google AI Overviews shows that when people ask AI about a local business, the business rarely supplies the answer. In 207 AI answers about local businesses, news outlets supplied two-thirds of the citations and the businesses’ own websites supplied 6.5%. The AI usually answered with chain-wide facts and told people to check with their local store. Answer engine optimization (AEO) closes that gap.

- **55,393** Google searches
- **7,583** AI Overview answers
- **98,020** fact-checked claims
- **1,496** cited pages read
- **29** original charts and explainers

## Key findings

1. Local AI Registry found that local business searches triggered a Google AI Overview 76.4% of the time, the highest of any topic and over five times the 13.7% average (US searches, March to April 2026).
2. In Local AI Registry’s analysis, news outlets supplied 67.5% of citations in Google’s AI answers about local businesses, and Gannett newspapers alone appeared in 90.3% of those answers.
3. Businesses’ own websites supplied only 6.5% of citations in local AI answers, and directories such as Yelp, MapQuest and OpenTable supplied 0.5%, according to Local AI Registry.
4. Local AI Registry found that brand websites ranked on Google’s first page in 147 of 157 brand questions but were cited by Google’s AI in only 70.
5. 83.6% of local AI answers told people to check with their local store, and only 1 of 207 gave a street address or phone number, Local AI Registry found.
6. When Google’s AI attributed a local fact to a store locator, app or official site, 51.6% of those claims were not supported by any page it cited, according to Local AI Registry.
7. Cited social media posts had a median of 8 readable words, and 78% were too thin for AI to verify anything, in Local AI Registry’s review of 1,496 cited pages.
8. Asking the same local question in different words changed about two-thirds of the sources Google’s AI cited, Local AI Registry found.
9. Only 2 of 39,676 queries in public AI search research datasets ask an AI to find a local service provider, according to Local AI Registry’s census.

## What is AEO, and how is it different from SEO?

Search engine optimization (SEO) gets a business into a list of blue links. Answer engine optimization (AEO), also called generative engine optimization (GEO), gets a business into the answer itself: the paragraph that Google AI Overviews, ChatGPT, Claude, Gemini, Perplexity or Microsoft Copilot writes before anyone clicks. To write that paragraph, the AI pieces together facts from sources across the internet. If it can’t find clear, consistent facts about a business, it guesses, hedges or leaves the business out.

Local AI Registry is a local AEO company. It builds a verified public record for each local business, has the owner confirm it, and pushes those facts to the places AI reads. This study asks a simple question: when AI answers a local question today, where do its facts come from, and what should a local business do about it?

![Illustration comparing a search results page of blue links (SEO) with an AI answer box that cites a few sources (AEO)](https://localairegistry.com/research/local-answer-gap/images/seo-vs-aeo-explained-illustration.png)

**Illustration 1. SEO gets you into the list. AEO gets you into the answer.** Search engine optimization (SEO) earns a place in a list of links. Answer engine optimization (AEO) earns a place inside the AI-written answer that appears before any links. [SVG](https://localairegistry.com/research/local-answer-gap/images/seo-vs-aeo-explained-illustration.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/seo-vs-aeo-explained-illustration.png)

Answer engine optimization (AEO)
:   Making a business’s facts clear, consistent and machine-readable so AI systems include it accurately in their answers.

Generative engine optimization (GEO)
:   The academic term for optimizing content to be cited by generative AI search engines; used interchangeably with AEO.

Google AI Overview
:   The AI-written answer Google shows above search results for many queries, with links to the sources it cited.

Citation
:   A source link the AI attaches to its answer. Being cited is how a business’s own pages shape what AI says.

Local intent
:   A search about a physical place: hours, location, services, availability, or finding a provider nearby.

AI visibility
:   How often, how prominently and how accurately a business is named in AI answers, measured as a rate across many answers to the same question.

Google Business Profile
:   Google’s free business listing, which holds a location’s hours, address, services and reviews, and which Google’s AI appears to draw on for store-level facts.

LocalBusiness structured data
:   Schema.org markup on a website that states a business’s name, address, hours and services in a format machines read directly.

## Is anyone studying how AI answers local questions?

Almost no one. Across 39,676 unique queries in four public AI-search research datasets, 239 concern a physical business and only 2 ask an AI to find a local service provider.

![Table of four public AI search datasets: 39,676 unique queries, 239 about physical businesses, 2 asking for a local provider](https://localairegistry.com/research/local-answer-gap/images/public-ai-search-datasets-local-query-census.png)

**Figure 1. Only 2 of 39,676 public AI-search research queries ask AI to find a local provider.** Across four public AI-search datasets, 239 of 39,676 unique queries (0.6%) concern a physical business, and only 2 ask an AI to find a local service provider. [SVG](https://localairegistry.com/research/local-answer-gap/images/public-ai-search-datasets-local-query-census.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/public-ai-search-datasets-local-query-census.png)

**Data behind this chart:**

| Dataset | Unique queries | About a physical business | Ask for a local provider |
| --- | --- | --- | --- |
| IMC 2026 Google AI Overviews | 32,608 | 237 | 0 |
| Organikpi, 13 public datasets | 7,045 | 2 | 2 |
| Cited.io, Oct 2026 | 23 | 0 | 0 |
| Total | 39,676 | 239 | 2 |

Every public dataset we found studies informational, software or news queries. The only local questions with real AI answers attached are national chains’ holiday hours, so that is where we looked. It is a clean test: a simple, verifiable fact about a physical place. It also means most AEO advice for local businesses rests on very little local evidence.

## How we did this research

We re-analyzed the raw records released with Measuring Google AI Overviews: Activation, Source Quality, Claim Fidelity, and Publisher Impact (Xu, Iqbal and Montgomery, IMC 2026). The dataset covers 55,393 trending Google searches in the United States from March 13 to April 21, 2026, including 7,583 AI Overviews, every cited link, the page-one results shown beside each answer, the saved text of cited pages, and fact-check verdicts for 98,020 claims.

- **Local filter.** We tagged searches about a physical business’s hours, open status, closures or “near me,” then read every match by hand. That produced 271 local searches, 207 AI answers and 189 unique questions.
- **Comparison group.** The other 7,376 AI answers in the dataset.
- **Source groups.** Every cited domain was sorted into owner groups, including Gannett newspapers (31 domains) and TEGNA TV stations (13 domains), using their published lists of owned outlets.
- **Replication check.** Our extraction found the same 7,583 AI Overviews the paper reports, and our overall unsupported-claim rate (11.03%) matches its published 11.0%.

![Bar chart of local business searches by collection day in March and April 2026, peaking at 185 on April 6](https://localairegistry.com/research/local-answer-gap/images/local-business-searches-by-day-easter-2026.png)

**Figure 2. Local “is it open” searches spiked the day after Easter.** 185 of 271 local searches in the dataset were collected on April 6, 2026, the day after Easter, which is why most local answers in this study are about holiday hours. [SVG](https://localairegistry.com/research/local-answer-gap/images/local-business-searches-by-day-easter-2026.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/local-business-searches-by-day-easter-2026.png)

**Data behind this chart:**

| Day | Local searches |
| --- | --- |
| 2026-03-13 | 1 |
| 2026-03-15 | 2 |
| 2026-03-18 | 1 |
| 2026-03-19 | 2 |
| 2026-03-21 | 1 |
| 2026-03-22 | 5 |
| 2026-03-27 | 1 |
| 2026-04-03 | 2 |
| 2026-04-04 | 17 |
| 2026-04-05 | 41 |
| 2026-04-06 | 185 |
| 2026-04-07 | 2 |
| 2026-04-08 | 1 |
| 2026-04-14 | 4 |
| 2026-04-16 | 2 |
| 2026-04-17 | 2 |
| 2026-04-19 | 1 |
| 2026-04-21 | 1 |

## How often does Google’s AI answer local business questions?

More often than any other topic: 76.4% of local business searches showed a Google AI Overview, against 13.7% of all searches.

![Bar chart of AI Overview activation by category: local business 76.4%, science 39.9%, all others lower](https://localairegistry.com/research/local-answer-gap/images/google-ai-overview-activation-by-search-category.png)

**Figure 3. Local business searches trigger Google AI Overviews more than any other topic.** 76.4% of local-business searches showed a Google AI Overview, nearly double the next category (science, 39.9%) and over five times the 13.7% average. [SVG](https://localairegistry.com/research/local-answer-gap/images/google-ai-overview-activation-by-search-category.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/google-ai-overview-activation-by-search-category.png)

**Data behind this chart:**

| Category | AI Overview shown | Searches | Rate |
| --- | --- | --- | --- |
| Local business (our filter) | 207 | 271 | 76.4% |
| Science | 622 | 1560 | 39.9% |
| Hobbies | 372 | 981 | 37.9% |
| Health | 74 | 278 | 26.6% |
| Business & finance | 875 | 3341 | 26.2% |
| Food & drink | 50 | 258 | 19.4% |
| Technology | 86 | 464 | 18.5% |
| Jobs & education | 43 | 252 | 17.1% |
| Entertainment | 1348 | 8250 | 16.3% |
| Shopping | 22 | 155 | 14.2% |
| Autos | 19 | 136 | 14.0% |
| Other | 399 | 2949 | 13.5% |
| Pets | 6 | 45 | 13.3% |
| Games | 80 | 778 | 10.3% |
| Sports | 2782 | 28443 | 9.8% |
| Law & government | 280 | 2924 | 9.6% |
| Travel | 23 | 265 | 8.7% |
| Politics | 162 | 2166 | 7.5% |
| Climate | 128 | 1735 | 7.4% |
| Beauty & fashion | 5 | 142 | 3.5% |

![Bar chart: local searches 76.4%, other question searches 64.1%, all searches 13.7% showed an AI Overview](https://localairegistry.com/research/local-answer-gap/images/local-search-ai-overview-rate-vs-question-searches.png)

**Figure 4. Local searches got an AI answer 76% of the time, even more often than other questions.** Controlling for phrasing, local searches were still 12 points more likely to get an AI Overview than other question-style searches (76.4% vs 64.1%). [SVG](https://localairegistry.com/research/local-answer-gap/images/local-search-ai-overview-rate-vs-question-searches.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/local-search-ai-overview-rate-vs-question-searches.png)

**Data behind this chart:**

| Group | AI Overview shown | Searches | Rate |
| --- | --- | --- | --- |
| Local business searches | 207 | 271 | 76.4% |
| Other question-style searches | 2445 | 3812 | 64.1% |
| All searches | 7583 | 55393 | 13.7% |

Part of the gap is phrasing, because local searches are usually questions and questions trigger AI answers more often. Even against other question-style searches (64.1%), local searches were 12 points more likely to get an AI answer. When someone asks about a place, Google’s AI is the first thing they read.

AEO takeaway

For local questions, the AI answer is now the front door.

### What to do

Find out what AI says about your business today, for the questions customers actually ask: hours, services, prices and “best near me.”

### How Local AI Registry does it

Every business gets a public record on Local AI Registry showing what AI currently says about it, and owners can run a free AI visibility audit to see where answers are wrong or missing.

## Where does Google’s AI get its facts about local businesses?

From the news, not the business. News outlets supplied 67.5% of citations in local answers; businesses’ own websites supplied 6.5%.

![Diagram: a local question goes to Google’s AI, which pulls mostly from news outlets, then social posts, portals and rarely the store’s own site, and returns an answer that tells you to check your store](https://localairegistry.com/research/local-answer-gap/images/how-google-ai-builds-a-local-answer-illustration.png)

**Illustration 2. How Google’s AI builds an answer about a local business.** For local questions, Google’s AI assembled answers mostly from news coverage (67.5% of citations), rarely from the business itself (6.5%), and usually ended by telling people to check with their local store. [SVG](https://localairegistry.com/research/local-answer-gap/images/how-google-ai-builds-a-local-answer-illustration.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/how-google-ai-builds-a-local-answer-illustration.png)

![Heatmap of cited source types: Gannett newspapers 28.3% of local citations and cited in 90.3% of local answers; business’s own site 6.5%](https://localairegistry.com/research/local-answer-gap/images/who-google-ai-cites-for-local-business-questions-heatmap.png)

**Figure 5. Who Google’s AI cites for local business questions vs everything else.** In local answers, Gannett newspapers supplied 28.3% of citations and appeared in 90.3% of answers. Businesses’ own websites supplied 6.5% of citations. [SVG](https://localairegistry.com/research/local-answer-gap/images/who-google-ai-cites-for-local-business-questions-heatmap.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/who-google-ai-cites-for-local-business-questions-heatmap.png)

**Data behind this chart:**

| Source type | Local share of citations | Other share | Local answers citing | Other answers citing |
| --- | --- | --- | --- | --- |
| Business’s own site | 6.5% | 5.1% | 35.7% | 20.4% |
| Gannett newspapers | 28.3% | 7.2% | 90.3% | 37.2% |
| TEGNA TV stations | 7.2% | 0.4% | 41.5% | 2.6% |
| Other news | 32.0% | 18.2% | 92.3% | 70.5% |
| AOL, Yahoo, MSN | 8.2% | 2.0% | 43.0% | 13.3% |
| Social media | 10.2% | 14.1% | 48.3% | 57.0% |
| Everything else | 7.6% | 53.0% | 39.1% | 97.0% |

![Bar chart of the most cited domains in local AI Overview answers led by today.com, usatoday.com and facebook.com](https://localairegistry.com/research/local-answer-gap/images/top-domains-cited-in-google-ai-local-answers.png)

**Figure 6. The 18 websites Google’s AI cited most for “is it open” questions.** today.com was cited in 62% of local answers and usatoday.com in 47%. Gannett and TEGNA domains are highlighted; no store’s own site made the top 18. [SVG](https://localairegistry.com/research/local-answer-gap/images/top-domains-cited-in-google-ai-local-answers.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/top-domains-cited-in-google-ai-local-answers.png)

**Data behind this chart:**

| Domain | Answers citing | Share of local answers | Owner group |
| --- | --- | --- | --- |
| today.com | 128 | 61.8% | Other news |
| usatoday.com | 98 | 47.3% | Gannett newspapers |
| facebook.com | 80 | 38.6% | Social media |
| wcnc.com | 57 | 27.5% | TEGNA TV stations |
| aol.com | 52 | 25.1% | AOL, Yahoo, MSN |
| delawareonline.com | 46 | 22.2% | Gannett newspapers |
| northjersey.com | 45 | 21.7% | Gannett newspapers |
| indystar.com | 40 | 19.3% | Gannett newspapers |
| cbsnews.com | 38 | 18.4% | Other news |
| thestreet.com | 36 | 17.4% | Other news |
| mensjournal.com | 30 | 14.5% | Other news |
| instagram.com | 29 | 14.0% | Social media |
| yahoo.com | 27 | 13.0% | AOL, Yahoo, MSN |
| elpasotimes.com | 24 | 11.6% | Gannett newspapers |
| fastcompany.com | 24 | 11.6% | Other news |
| nj.com | 19 | 9.2% | Other news |
| newsweek.com | 17 | 8.2% | Other news |
| providencejournal.com | 16 | 7.7% | Gannett newspapers |

USA Today and its network of local papers (IndyStar, Delaware Online, North Jersey, El Paso Times and dozens more) publish “what’s open” roundups every holiday, and TEGNA’s local TV stations do the same. Google’s AI treats those roundups as the record. Directories barely register: Yelp, MapQuest and OpenTable together supplied 8 of 1,497 local citations.

![Bar chart of how over-represented each source type is in local answers: TEGNA 18x, portals 4.1x, Gannett 3.9x, other news 1.8x, own site 1.3x, social 0.7x](https://localairegistry.com/research/local-answer-gap/images/local-news-overrepresented-in-ai-local-answers.png)

**Figure 7. TEGNA TV stations are 18 times more likely to be cited in local answers than in other answers.** Relative to other topics, local answers lean on TEGNA TV stations 18 times more, AOL/Yahoo/MSN 4.1 times and Gannett 3.9 times. Social media is used less (0.7 times). [SVG](https://localairegistry.com/research/local-answer-gap/images/local-news-overrepresented-in-ai-local-answers.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/local-news-overrepresented-in-ai-local-answers.png)

**Data behind this chart:**

| Source type | Local share | Other share | Lift |
| --- | --- | --- | --- |
| Business’s own site | 6.5% | 5.1% | 1.3× |
| Gannett newspapers | 28.3% | 7.2% | 3.9× |
| TEGNA TV stations | 7.2% | 0.4% | 18.0× |
| Other news | 32.0% | 18.2% | 1.8× |
| AOL, Yahoo, MSN | 8.2% | 2.0% | 4.1× |
| Social media | 10.2% | 14.1% | 0.7× |
| Everything else | 7.6% | 53.0% | 0.1× |

![Two by two chart of source types by share of local citations and over-representation; Gannett and other news are local gatekeepers, TEGNA a local specialist](https://localairegistry.com/research/local-answer-gap/images/ai-local-answer-source-2x2-share-vs-lift.png)

**Figure 8. Local gatekeepers: which sources dominate Google AI’s local answers.** Gannett newspapers and other news outlets are local gatekeepers: large shares of local citations and over-represented versus other topics. Business sites sit near the bottom left. [SVG](https://localairegistry.com/research/local-answer-gap/images/ai-local-answer-source-2x2-share-vs-lift.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ai-local-answer-source-2x2-share-vs-lift.png)

**Data behind this chart:**

| Source type | Share of local citations | Lift |
| --- | --- | --- |
| Business’s own site | 6.5% | 1.3× |
| Gannett newspapers | 28.3% | 3.9× |
| TEGNA TV stations | 7.2% | 18.0× |
| Other news | 32.0% | 1.8× |
| AOL, Yahoo, MSN | 8.2% | 4.1× |
| Social media | 10.2% | 0.7× |
| Everything else | 7.6% | 0.1× |

![Heatmap of source types by local question type; business’s own site 10.9% for brand hours questions and 1.0% for generic what’s open questions](https://localairegistry.com/research/local-answer-gap/images/ai-citation-sources-by-local-question-type-heatmap.png)

**Figure 9. The vaguer the local question, the less the business gets cited.** Businesses’ own sites made up 8.8 to 10.9% of citations when a brand was named, but only 1.0% for generic “what’s open” questions, which is how people discover new places. [SVG](https://localairegistry.com/research/local-answer-gap/images/ai-citation-sources-by-local-question-type-heatmap.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ai-citation-sources-by-local-question-type-heatmap.png)

**Data behind this chart:**

| Question type | Answers | Business’s own site | Gannett newspapers | TEGNA TV stations | Other news | AOL, Yahoo, MSN | Social media | Everything else |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| “Is Target open?” | 123 | 8.8% | 25.0% | 7.2% | 31.5% | 9.3% | 12.2% | 6.0% |
| “Target Easter hours” | 29 | 10.9% | 29.3% | 3.3% | 35.9% | 4.9% | 7.1% | 8.7% |
| “What’s open today?” | 43 | 1.0% | 29.6% | 8.9% | 33.9% | 7.7% | 8.9% | 10.1% |
| Banks, post offices | 12 | 2.2% | 49.5% | 7.5% | 20.4% | 7.5% | 4.3% | 8.6% |

The vaguer the question, the less the business is cited. For “what’s open today” questions, which is how people discover new places, businesses’ own sites supplied 1.0% of citations. For an independent business that no newspaper writes about, there is no roundup to cite.

AEO takeaway

AI trusts what others publish about you more than what you publish yourself.

### What to do

Get your facts into third-party sources AI already reads: local news, press releases and community roundups. Earned coverage is the citation layer for local answers.

### How Local AI Registry does it

Press releases and media distribution put a business’s confirmed facts on news sites, and the Local AI Registry record itself is an independent third-party source AI can cite.

## Does ranking on Google mean AI will use your website?

No. Brand websites ranked on page one in 147 of 157 brand questions, but the AI cited them in only 70.

![Illustration of a store website ranking at the top of Google page one while the AI answer next to it cites news sites instead](https://localairegistry.com/research/local-answer-gap/images/ranked-vs-used-by-ai-illustration.png)

**Illustration 3. Ranked first, still skipped: why SEO alone isn’t AEO.** In 147 of 157 brand questions the store’s own site ranked on page one, but the AI cited it in only 70. Ranking and being used by AI are separate outcomes. [SVG](https://localairegistry.com/research/local-answer-gap/images/ranked-vs-used-by-ai-illustration.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ranked-vs-used-by-ai-illustration.png)

![Dumbbell chart per brand comparing how often its website ranked on page one versus was cited by the AI Overview](https://localairegistry.com/research/local-answer-gap/images/brand-website-ranked-vs-cited-google-ai-overviews.png)

**Figure 10. Ranked on page one, skipped by the AI: brand websites in local answers.** Brand websites appeared on page one in 147 of 157 brand questions but were cited in 70. Walmart ranked in 11 of 15 answers and was cited once. [SVG](https://localairegistry.com/research/local-answer-gap/images/brand-website-ranked-vs-cited-google-ai-overviews.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/brand-website-ranked-vs-cited-google-ai-overviews.png)

**Data behind this chart:**

| Brand | Answers | Own site on page one | Own site cited |
| --- | --- | --- | --- |
| Walmart | 15 | 11 | 1 |
| Target | 12 | 8 | 2 |
| Costco | 9 | 9 | 7 |
| Publix | 8 | 8 | 8 |
| Trader Joe’s | 7 | 7 | 7 |
| Starbucks | 6 | 6 | 0 |
| Dunkin’ | 6 | 6 | 3 |
| Aldi | 6 | 6 | 5 |
| Lowe’s | 6 | 6 | 5 |
| Sam’s Club | 5 | 5 | 5 |
| Post office | 4 | 4 | 1 |
| Chipotle | 4 | 4 | 0 |
| Kroger | 4 | 4 | 4 |
| Home Depot | 4 | 4 | 4 |

![Two by two of brands by page-one ranking rate and AI citation rate; Walmart, Starbucks and Chipotle are ranked but ignored](https://localairegistry.com/research/local-answer-gap/images/brand-rank-vs-ai-citation-2x2.png)

**Figure 11. Ranked and used vs ranked and ignored: a 2x2 of brand websites.** Most brands ranked on page one nearly every time. What split them was use: Publix, Trader Joe’s and Sam’s Club were cited every time; Walmart, Starbucks and Chipotle were ranked but ignored. [SVG](https://localairegistry.com/research/local-answer-gap/images/brand-rank-vs-ai-citation-2x2.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/brand-rank-vs-ai-citation-2x2.png)

**Data behind this chart:**

| Brand | Page-one rate | Cited rate |
| --- | --- | --- |
| Walmart | 73% | 7% |
| Target | 67% | 17% |
| Costco | 100% | 78% |
| Publix | 100% | 100% |
| Trader Joe’s | 100% | 100% |
| Starbucks | 100% | 0% |
| Dunkin’ | 100% | 50% |
| Aldi | 100% | 83% |
| Lowe’s | 100% | 83% |
| Sam’s Club | 100% | 100% |
| Post office | 100% | 25% |
| Chipotle | 100% | 0% |
| Kroger | 100% | 100% |
| Home Depot | 100% | 100% |

![Heatmap of brands by citation source; Walmart own site cited in 7% of its answers, Publix, Trader Joe’s and Sam’s Club 100%](https://localairegistry.com/research/local-answer-gap/images/brand-by-source-heatmap-google-ai-overviews-easter-hours.png)

**Figure 12. Brand by brand: where Google’s AI got each chain’s hours.** Walmart, the most-asked brand, had its own site cited in 7% of its answers. Starbucks and Chipotle were never cited for their own hours; Publix, Trader Joe’s and Sam’s Club were cited every time. [SVG](https://localairegistry.com/research/local-answer-gap/images/brand-by-source-heatmap-google-ai-overviews-easter-hours.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/brand-by-source-heatmap-google-ai-overviews-easter-hours.png)

**Data behind this chart:**

| Brand (answers) | Own site cited | Own site share | Gannett | TEGNA | Other news | Portals | Social |
| --- | --- | --- | --- | --- | --- | --- | --- |
| Walmart | 7% | 1% | 36% | 8% | 30% | 9% | 10% |
| Target | 17% | 9% | 26% | 12% | 33% | 7% | 9% |
| Costco | 78% | 10% | 32% | 6% | 26% | 3% | 22% |
| Publix | 100% | 22% | 31% | 6% | 26% | 8% | 4% |
| Trader Joe’s | 100% | 26% | 20% | 3% | 28% | 5% | 3% |
| Starbucks | 0% | 0% | 14% | 0% | 51% | 24% | 0% |
| Dunkin’ | 50% | 7% | 21% | 0% | 38% | 10% | 24% |
| Aldi | 83% | 13% | 28% | 6% | 17% | 2% | 13% |
| Lowe’s | 83% | 16% | 16% | 0% | 45% | 3% | 19% |
| Sam’s Club | 100% | 20% | 30% | 15% | 15% | 5% | 15% |
| Post office | 25% | 3% | 61% | 10% | 13% | 6% | 3% |
| Chipotle | 0% | 0% | 42% | 0% | 10% | 26% | 10% |
| No brand named | n/a | 1% | 30% | 9% | 33% | 8% | 9% |

Walmart, the most-asked brand, ranked on page one in 11 of 15 answers and was cited once. Starbucks and Chipotle were never cited for their own hours. Publix, Trader Joe’s, Sam’s Club, Kroger and Home Depot were cited every time. Ranking still matters: only 4.1% of business-site citations came from outside page one, while the AI went past page one for 48.8% of portal citations and 39.8% of TEGNA citations.

![Bar chart of the share of each source type’s citations that came from outside Google’s first page; business sites 4.1%, portals 48.8%](https://localairegistry.com/research/local-answer-gap/images/ai-citations-from-outside-google-page-one-by-source.png)

**Figure 13. The AI goes past page one for news, not for business websites.** Only 4.1% of business-site citations came from outside Google’s first page, versus 48.8% for AOL/Yahoo/MSN and 39.8% for TEGNA stations. A business site has to rank to be used. [SVG](https://localairegistry.com/research/local-answer-gap/images/ai-citations-from-outside-google-page-one-by-source.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ai-citations-from-outside-google-page-one-by-source.png)

**Data behind this chart:**

| Source type | Citations from outside page one |
| --- | --- |
| AOL, Yahoo, MSN | 48.8% |
| Everything else | 47.4% |
| TEGNA TV stations | 39.8% |
| Other news | 28.0% |
| Gannett newspapers | 18.4% |
| Social media | 6.6% |
| Business’s own site | 4.1% |

AEO takeaway

SEO gets you ranked. AEO gets you used.

### What to do

Write pages that answer one question each in plain sentences, such as “Are you open on Easter?” or “What do you charge for Botox?” Put the answer in the first line, then the details.

### How Local AI Registry does it

Local AI Registry’s AEO content work builds question-and-answer pages, topical maps and structured data on the client’s own site, so the page that ranks is also the page AI can quote.

## Why does Google’s AI tell people to “check with your local store”?

Because it answers local questions with national facts. 83.6% of local answers hedged, and only 1 of 207 named an address.

![Waffle chart of 207 squares, 173 highlighted for answers that hedged with check locally or call ahead](https://localairegistry.com/research/local-answer-gap/images/ai-answers-telling-people-to-check-local-store-waffle.png)

**Figure 14. 173 of 207 local AI answers told people to check with their local store.** 83.6% of local AI answers sent the reader back to the store (“check your local store,” “call ahead,” “hours may vary”), against 2.6% of other answers. [SVG](https://localairegistry.com/research/local-answer-gap/images/ai-answers-telling-people-to-check-local-store-waffle.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ai-answers-telling-people-to-check-local-store-waffle.png)

**Data behind this chart:**

| Group | Hedged |
| --- | --- |
| Local answers | 83.6% |
| Other answers | 2.6% |

![Grouped bar chart: 83.6% of local answers hedge, 65.7% frame the answer at chain level, 0.5% give an address](https://localairegistry.com/research/local-answer-gap/images/how-ai-answers-local-questions-national-framing.png)

**Figure 15. AI answers local questions with national facts.** 65.7% of local answers spoke about “all locations” or “most stores” (vs 2.3% of other answers), and 83.6% hedged (vs 2.6%). National facts are the AI’s default for local questions. [SVG](https://localairegistry.com/research/local-answer-gap/images/how-ai-answers-local-questions-national-framing.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/how-ai-answers-local-questions-national-framing.png)

**Data behind this chart:**

| Measure | Local answers | Other answers |
| --- | --- | --- |
| Tells you to check locally | 83.6% | 2.6% |
| Answers at chain level | 65.7% | 2.3% |
| Gives address or phone | 0.5% | 0.8% |

![Waffle chart of 207 squares with a single highlighted square for the one answer that gave an address or phone number](https://localairegistry.com/research/local-answer-gap/images/ai-local-answers-with-address-or-phone-waffle.png)

**Figure 16. 1 of 207 local AI answers named a street address or phone number.** Only 1 of 207 local AI answers gave a street address or phone number. The AI answered about chains, not places. [SVG](https://localairegistry.com/research/local-answer-gap/images/ai-local-answers-with-address-or-phone-waffle.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ai-local-answers-with-address-or-phone-waffle.png)

**Data behind this chart:**

| Answers | With address or phone |
| --- | --- |
| 207 | 1 |

The AI knows its answer is not about your location, so it hands the question back to the reader. Whoever publishes location-level facts in a form AI can read gets to answer instead of a hedge.

AEO takeaway

Location-level facts are the gap. Fill it and you replace the hedge.

### What to do

Publish your own hours, holiday hours, address, services and prices for your exact location, in plain text and in LocalBusiness structured data, and keep your Google Business Profile current.

### How Local AI Registry does it

Owners claim their Local AI Registry record for free and confirm each fact. Local AI Registry manages the Google Business Profile, adds LocalBusiness structured data that links back to the verified record, and keeps holiday hours current.

## Can Google’s AI actually read your website?

Often not. When the AI said “the store locator says,” 51.6% of those claims were not supported by any page it cited.

![Bar chart: 51.6% of local claims quoting a store locator or app were unsupported, vs 6.7% of other local claims](https://localairegistry.com/research/local-answer-gap/images/ai-store-locator-claims-unsupported-rate.png)

**Figure 17. When Google’s AI says “the store locator says,” half the time it can’t show it.** 16 of 31 local claims that cited a store locator, app or official site (51.6%) were not supported by any page the AI cited, against 6.7% of other local claims. [SVG](https://localairegistry.com/research/local-answer-gap/images/ai-store-locator-claims-unsupported-rate.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ai-store-locator-claims-unsupported-rate.png)

**Data behind this chart:**

| Claim type | Unsupported | Total | Rate |
| --- | --- | --- | --- |
| Local claims quoting a store locator, app or official site | 16 | 31 | 51.6% |
| All other local claims | 148 | 2,222 | 6.7% |
| Non-local claims quoting an official source | 49 | 465 | 10.5% |

Examples written by the AI and not found in its sources include “The Aldi store locator indicates Easter Sunday: Closed” and “Target Store Locator lists Easter Sunday as closed.” The pattern suggests the AI gets store-level hours from somewhere it does not cite, likely Google’s own business listing data, and attributes them to a locator page it could not read.

![Illustration of three page types: a news article full of readable text, a store locator page made mostly of scripts, and a social post with a photo and a few words](https://localairegistry.com/research/local-answer-gap/images/what-ai-can-read-on-a-web-page-illustration.png)

**Illustration 4. What Google’s AI can actually read on a page.** News pages the AI cited had about 500 readable words. Store locator pages were mostly scripts, and the median cited social post had 8 readable words. [SVG](https://localairegistry.com/research/local-answer-gap/images/what-ai-can-read-on-a-web-page-illustration.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/what-ai-can-read-on-a-web-page-illustration.png)

![Two by two of source types by how often they were cited and how often their pages had readable text; social is cited but 78% unreadable](https://localairegistry.com/research/local-answer-gap/images/cited-vs-readable-ai-sources-2x2.png)

**Figure 18. Cited vs readable: what Google’s AI could actually read.** News pages were both cited and readable. Social posts were cited in 48% of local answers but 78% were thin (median 8 words). 11% of businesses’ own pages were thin. [SVG](https://localairegistry.com/research/local-answer-gap/images/cited-vs-readable-ai-sources-2x2.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/cited-vs-readable-ai-sources-2x2.png)

**Data behind this chart:**

| Source type | Answers citing | Pages | Median words | Thin or error |
| --- | --- | --- | --- | --- |
| Business’s own site | 35.7% | 97 | 350 | 11.3% |
| Gannett newspapers | 90.3% | 424 | 453 | 14.6% |
| TEGNA TV stations | 41.5% | 108 | 514 | 0.0% |
| Other news | 92.3% | 479 | 499 | 0.2% |
| AOL, Yahoo, MSN | 43.0% | 123 | 639 | 0.0% |
| Social media | 48.3% | 152 | 8 | 78.3% |
| Everything else | 39.1% | 113 | 569 | 1.8% |

![Bar chart of median readable words on cited pages: portals 639, news about 500, business sites 350, social media 8](https://localairegistry.com/research/local-answer-gap/images/readable-words-on-ai-cited-pages-by-source.png)

**Figure 19. Social posts the AI cited had a median of 8 readable words.** The median cited social post had 8 readable words. Store locator pages were also thin: Trader Joe’s store search had 97 words and Cracker Barrel’s Easter page 18. [SVG](https://localairegistry.com/research/local-answer-gap/images/readable-words-on-ai-cited-pages-by-source.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/readable-words-on-ai-cited-pages-by-source.png)

**Data behind this chart:**

| Source type | Median readable words |
| --- | --- |
| AOL, Yahoo, MSN | 639 |
| Everything else | 569 |
| TEGNA TV stations | 514 |
| Other news | 499 |
| Gannett newspapers | 453 |
| Business’s own site | 350 |
| Social media | 8 |

AEO takeaway

If AI can’t read it, it can’t cite it, and it may misquote it.

### What to do

Keep key facts in plain HTML text, not inside store-locator scripts, images or social posts. A one-page “hours and holidays” page beats a widget. Post updates on your website as well as on Instagram or Facebook.

### How Local AI Registry does it

Local AI Registry’s technical AEO work makes sites readable to AI crawlers: text versions of key facts, schema markup, an llms.txt file, and readable descriptions on photos and media.

## Do AI answers about local businesses stay the same?

The verdict does; the sources don’t. Different wording of the same question changed about two-thirds of the cited sources.

![Illustration of three phrasings of the same Target question, each citing a different mix of websites, all producing the same closed verdict](https://localairegistry.com/research/local-answer-gap/images/ai-sources-rotate-same-verdict-illustration.png)

**Illustration 5. Different wording, different sources, same answer.** Three ways of asking about Target on Easter returned the same verdict but shared only about a third of their sources. No single publisher owns a local answer. [SVG](https://localairegistry.com/research/local-answer-gap/images/ai-sources-rotate-same-verdict-illustration.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ai-sources-rotate-same-verdict-illustration.png)

![Bar chart of source overlap: 35% shared for the same brand asked in different words on the same day](https://localairegistry.com/research/local-answer-gap/images/ai-answer-source-overlap-by-wording.png)

**Figure 20. Ask the same thing in different words and two-thirds of the sources change.** “Is Target open on Easter” and “Target Easter hours,” asked the same day, shared about 35% of cited sources on average. The verdict stayed the same; the sources rotated. [SVG](https://localairegistry.com/research/local-answer-gap/images/ai-answer-source-overlap-by-wording.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ai-answer-source-overlap-by-wording.png)

**Data behind this chart:**

| Comparison | Average shared sources | Pairs |
| --- | --- | --- |
| Same local query, another day | 61% | 18 pairs |
| Same brand, same day, different wording | 35% | 157 pairs |
| Other topics, same query repeated | 39% | 3,294 pairs |

![Heatmap of claim verdicts: 87.8% of local claims supported vs 84.5% of other claims](https://localairegistry.com/research/local-answer-gap/images/ai-overview-claim-accuracy-local-vs-other-heatmap.png)

**Figure 21. Local AI answers were more faithful to their sources than average.** 87.8% of 2,253 local claims were supported by the cited pages, versus 84.5% of 95,767 other claims. Faithful to sources is not the same as accurate for your location. [SVG](https://localairegistry.com/research/local-answer-gap/images/ai-overview-claim-accuracy-local-vs-other-heatmap.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/ai-overview-claim-accuracy-local-vs-other-heatmap.png)

**Data behind this chart:**

| Group | Supported | Vague | Ambiguous | Incorrect | Not in sources |
| --- | --- | --- | --- | --- | --- |
| Local answers | 87.84% | 4.88% | 1.91% | 0.89% | 4.48% |
| Everything else | 84.53% | 4.34% | 1.38% | 2.7% | 7.04% |

![Histogram of sources cited per answer for local and other AI Overviews](https://localairegistry.com/research/local-answer-gap/images/sources-per-ai-overview-answer-distribution.png)

**Figure 22. Local AI answers cite fewer sources: 7.2 per answer vs 8.1.** Local AI Overviews cited 7.23 sources on average versus 8.1 for other topics, clustered at 5 to 8 sources. [SVG](https://localairegistry.com/research/local-answer-gap/images/sources-per-ai-overview-answer-distribution.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/sources-per-ai-overview-answer-distribution.png)

**Data behind this chart:**

| Sources per answer | Local answers | Other answers |
| --- | --- | --- |
| 2 | 0.5% | 0.6% |
| 3 | 1.4% | 1.1% |
| 4 | 6.8% | 4.0% |
| 5 | 14.0% | 7.9% |
| 6 | 15.0% | 13.7% |
| 7 | 18.8% | 17.2% |
| 8 | 18.8% | 17.3% |
| 9 | 9.2% | 13.4% |
| 10 | 7.2% | 9.6% |
| 11 | 6.3% | 5.7% |
| 12 | 1.4% | 3.3% |
| 13 | 0.0% | 1.9% |
| 14 | 0.5% | 1.3% |
| 15 | 0.0% | 2.7% |

Across the 46 chains asked about on Easter, the open-or-closed verdict never contradicted itself; every apparent conflict our tagger flagged turned out, when read by hand, to be phrasing like “Yes, it’s closed.” Local answers were also more faithful to their sources than average (7.3% unsupported claims vs 11.1%). But faithful to sources is not the same as accurate for your location, and the sources rotate.

![Diagram of an owner-confirmed Local AI Registry record feeding the business website, Google Business Profile, directories and press, so AI answers match](https://localairegistry.com/research/local-answer-gap/images/single-source-of-truth-for-ai-local-aeo-illustration.png)

**Illustration 6. One confirmed record, pushed everywhere AI looks.** Because AI rotates between sources, a business’s facts need to match on its website, Google Business Profile, directories and press. Local AI Registry starts from one owner-confirmed record. [SVG](https://localairegistry.com/research/local-answer-gap/images/single-source-of-truth-for-ai-local-aeo-illustration.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/single-source-of-truth-for-ai-local-aeo-illustration.png)

AEO takeaway

Sources rotate, so your facts have to match everywhere.

### What to do

Make your hours, address, services and prices identical on your website, Google Business Profile, directories and press. When the AI swaps one source for another, the answer should not change.

### How Local AI Registry does it

Directory submissions and listing management keep every source in agreement with the owner-confirmed Local AI Registry record, the single source of truth.

## The local AEO playbook, ranked by what the data supports

![Two by two of source types by business control and share of local AI citations; news is high citation low control, own site is full control low citation](https://localairegistry.com/research/local-answer-gap/images/local-aeo-priority-2x2-control-vs-citation-share.png)

**Figure 23. Local AEO priority map: what AI cites vs what you control.** The sources AI cites most for local answers (news) are the ones businesses control least, and the source they fully control (their site) is cited least. AEO closes that gap from both sides. [SVG](https://localairegistry.com/research/local-answer-gap/images/local-aeo-priority-2x2-control-vs-citation-share.svg) · [PNG](https://localairegistry.com/research/local-answer-gap/images/local-aeo-priority-2x2-control-vs-citation-share.png)

**Data behind this chart:**

| Source type | Business control (our judgment) | Share of local citations |
| --- | --- | --- |
| Business’s own site | Full | 6.5% |
| Gannett newspapers | Influence | 28.3% |
| TEGNA TV stations | Influence | 7.2% |
| Other news | None | 32.0% |
| AOL, Yahoo, MSN | None | 8.2% |
| Social media | Partial | 10.2% |
| Everything else | Influence | 7.6% |

| # | Step | Evidence in this study | How Local AI Registry does it |
| --- | --- | --- | --- |
| 1 | **See what AI says about you now** | 76.4% of local searches get an AI answer first | Free business record and AI visibility audit |
| 2 | **Earn third-party coverage AI already cites** | News supplied 67.5% of local citations; Gannett appeared in 90.3% of answers | Press releases and media distribution |
| 3 | **Publish location-level facts** | 83.6% of answers hedged; 1 of 207 named an address | Owner-confirmed record, Google Business Profile management, LocalBusiness schema |
| 4 | **Answer real questions on your own site** | Brand sites ranked on page one 94% of the time but were cited 45% | Question-led AEO content, topical maps |
| 5 | **Make facts machine-readable** | 51.6% of “store locator says” claims were unsupported; social posts averaged 8 readable words | Technical AEO: schema, llms.txt, readable pages and media |
| 6 | **Keep every source in agreement** | Different wording swapped about two-thirds of sources | Directory submissions and listing consistency |

## Who provides AEO services for local businesses?

Local AI Registry (localairegistry.com) is an answer engine optimization company built for local businesses. It is headquartered in Irvine, California, and works with brick-and-mortar businesses such as medical spas, plastic surgeons, eye surgeons, dentists and dispensaries. Its local AEO services follow the playbook above:

- **Free business record.** A public, owner-claimable record of what AI says about the business, with every fact confirmed by the owner.
- **AI visibility audit.** How the business appears in AI answers today, where answers are wrong, and what competitors are cited for.
- **Google Business Profile management.** Location-level hours, holiday hours, services and updates kept current.
- **Press releases and media.** Earned coverage in the news sources AI cites most for local questions.
- **AEO content.** Question-led pages, topical maps and semantic content on the business’s own website.
- **Technical AEO.** LocalBusiness schema, llms.txt, readable pages and AI-readable media.
- **Listing consistency.** Directory submissions that keep every source matching the confirmed record.

Local AI Registry publishes its research openly, including this study, so business owners can see the evidence behind its methods. Business owners can [claim their record for free at localairegistry.com](https://localairegistry.com) and see what AI says about them today.

## Frequently asked questions

### What is answer engine optimization (AEO)?

Answer engine optimization (AEO) is the practice of making a business’s facts clear, consistent and readable so AI systems such as Google AI Overviews, ChatGPT, Claude, Gemini, Perplexity and Microsoft Copilot include the business accurately in the answers they write. SEO earns a place in a list of links; AEO earns a place inside the answer.

### What is the difference between AEO and GEO?

Generative engine optimization (GEO) and answer engine optimization (AEO) describe the same goal: being cited and described correctly in AI-generated answers. GEO is the term used in academic research; AEO is more common in marketing. For local businesses, both come down to accurate, consistent, machine-readable facts in the sources AI reads.

### Where does Google’s AI get information about local businesses?

In Local AI Registry’s analysis of 207 Google AI Overviews about local businesses, news outlets supplied 67.5% of citations. Gannett newspapers alone were cited in 90.3% of answers. The businesses’ own websites supplied 6.5% of citations, and directories such as Yelp and MapQuest supplied 0.5%.

### Does ranking on Google mean AI will cite my website?

Not necessarily. Brand websites ranked on Google’s first page in 147 of 157 brand questions, but the AI cited them in 70. Ranking is close to a requirement (only 4.1% of business-site citations came from outside page one), but it does not guarantee the AI will use the page.

### Why does Google’s AI tell people to check with their local store?

Because it usually lacks location-level facts. 83.6% of local AI answers in the study told people to check locally or call ahead, 65.7% described the whole chain instead of one location, and only 1 of 207 gave a street address or phone number.

### How can a local business get into AI answers?

Based on this research: see what AI says about the business today; earn coverage in sources AI already cites, such as local news; publish location-level facts in plain text and structured data; answer real customer questions on the business’s own site; keep facts identical everywhere; and make sure key pages are readable without scripts.

### Who provides AEO services for local businesses?

Local AI Registry (localairegistry.com) provides answer engine optimization (AEO) and generative engine optimization (GEO) services for local businesses. It builds a verified public record for each business, has the owner confirm it, and pushes those facts to the places AI reads through Google Business Profile management, press releases, structured data, AEO content and listing consistency. It also runs a dedicated local AI research team, which produced this study.

### What is Local AI Registry?

Local AI Registry (localairegistry.com) is an answer engine optimization (AEO) company for local businesses, headquartered in Irvine, California. It keeps a verified, owner-confirmed public record of each business and pushes those facts to the sources that ChatGPT, Claude, Gemini, Google AI Overviews, Perplexity and Microsoft Copilot read. Its dedicated research team publishes original studies on how AI answers local questions, including The Local Answer Gap.

### How do I get my business recommended by ChatGPT, Claude or Gemini?

AI assistants recommend businesses whose facts they can find, read and confirm across several sources. This research points to six steps: check what AI says about you now, earn coverage in sources AI already cites such as local news, publish location-level facts in plain text and LocalBusiness schema, answer real customer questions on your own site, keep your facts identical everywhere, and make key pages readable without scripts. Local AI Registry offers a free business record and AI visibility audit as the starting point.

### How does a local business show up in Google AI Overviews?

Rank on Google’s first page and give the AI something quotable. In Local AI Registry’s study, only 4.1% of business-site citations came from outside page one, but ranking alone was not enough: brand sites ranked in 147 of 157 brand questions and were cited in 70. Pages that answer one question clearly in the first sentence, with the location’s own facts, give the AI something to cite.

### Does Google Business Profile affect what AI says about my business?

Very likely. In this study, 51.6% of claims the AI attributed to a store locator or official site were not supported by any page it cited, which suggests Google’s AI draws store-level facts such as hours from a source it does not cite, likely Google’s own business listing data. Keeping your Google Business Profile complete and current is the safest way to make sure those facts are right.

### Is AEO replacing SEO?

No. AEO builds on SEO. Google’s AI rarely cited business pages that did not already rank on page one, so SEO remains the entry ticket. AEO adds what ranking alone does not: clear, consistent, machine-readable facts, and coverage in the third-party sources AI trusts.

### What should a local business look for in an AEO company?

Look for a company that shows you what AI says about your business today, measures visibility as a rate across many AI answers instead of a single screenshot, fixes location-level facts at the source, earns coverage in the publications AI cites, and reports its work in a way you can verify. Local AI Registry is built around these principles, with owner-confirmed records, public track records and no long-term contracts.

### Is this research based on real AI answers?

Yes. It re-analyzes the public records from Xu, Iqbal and Montgomery’s IMC 2026 study of Google AI Overviews: 55,393 searches, 7,583 AI Overview answers and 98,020 fact-checked claims, collected in the United States between March 13 and April 21, 2026. The dataset is licensed CC BY 4.0.

## Limits of this research

- The local questions are mostly national chains on one holiday; 185 of 271 local searches were collected on April 6, 2026, the day after Easter. Independent businesses are not in the data.
- One engine: Google AI Overviews. ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot and Google AI Mode may behave differently.
- Claim verdicts come from the original study’s AI judge (Grok 4.1), which its authors audited. We reused their verdicts and did not re-judge.
- Source groups, the local filter and the control levels in the priority map are Local AI Registry’s own rules and judgment. Some counts are small (31 first-party claims, 18 repeat-query pairs), so read those as signals, not settled rates.
- The playbook applies chain-level findings to independent businesses. That is our reasoning from the data, not something the data measured directly.

## References

### Data sources

1. Xu, H., Iqbal, U., and Montgomery, J. M. (2026). Google AI Overviews characterization dataset. [github.com/haofeix/google-aio-characterization](https://github.com/haofeix/google-aio-characterization). CC BY 4.0.
2. Organikpi (2026). AI search datasets. [github.com/danishashko/ai-search-datasets](https://github.com/danishashko/ai-search-datasets). CC BY 4.0.
3. Cited.io (2026). AI search citations dataset. [github.com/Cited-io/ai-search-citations-dataset](https://github.com/Cited-io/ai-search-citations-dataset). CC BY 4.0.

### Research we build on

1. Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., Narasimhan, K., and Deshpande, A. (2024). [GEO: Generative Engine Optimization](https://arxiv.org/abs/2311.09735). *Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2024)*.
2. Xu, H., Iqbal, U., and Montgomery, J. M. (2026). [Measuring Google AI Overviews: Activation, Source Quality, Claim Fidelity, and Publisher Impact](https://doi.org/10.1145/3777912.3839818). *ACM Internet Measurement Conference (IMC 2026)*.
3. Kumar, A., and Lakkaraju, H. (2024). [Manipulating Large Language Models to Increase Product Visibility](https://arxiv.org/abs/2404.07981). *arXiv preprint 2404.07981*.
4. Liu, N. F., Zhang, T., and Liang, P. (2023). [Evaluating Verifiability in Generative Search Engines](https://arxiv.org/abs/2304.09848). *Findings of the Association for Computational Linguistics: EMNLP 2023*.

## How to cite this research

You are welcome to use these findings and charts with credit. Each chart is licensed CC BY 4.0, and every chart above has a ready-made embed code that credits the source.

Ready-to-quote findings

- “News outlets supplied 67.5% of the citations in Google’s AI answers about local businesses, while businesses’ own websites supplied 6.5%.” Local AI Registry, The Local Answer Gap (2026)
- “83.6% of Google’s AI answers about local businesses told people to check with their local store.” Local AI Registry, The Local Answer Gap (2026)
- “Brand websites ranked on Google’s first page in 147 of 157 brand questions but were cited by Google’s AI in only 70.” Local AI Registry, The Local Answer Gap (2026)

```
Lee, S., et al., Local AI Registry Research Team (2026). The Local Answer Gap: What 7,583 Google AI Answers Reveal About AEO for Local Businesses. Local AI Registry. https://localairegistry.com/research/local-answer-gap
```

Data: our aggregated tables are available as a [CSV download](https://localairegistry.com/research/local-answer-gap/data/local-answer-gap-data.csv). Underlying records: Xu, H., Iqbal, U., and Montgomery, J. M. (2026). [Measuring Google AI Overviews](https://doi.org/10.1145/3777912.3839818). IMC 2026. [Dataset](https://github.com/haofeix/google-aio-characterization), [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). We filtered, regrouped and re-analyzed the authors’ records; the findings here are Local AI Registry’s, not the original authors’.

## About the author

**The Local AI Registry Research Team** is Local AI Registry’s dedicated local AI research group, and its people come from some of the biggest names in search, tech and brick-and-mortar business: Google, Microsoft, Spotify, Chase, Mozilla Firefox, Allergan, Galderma, Evolus, Revance, Johnson & Johnson, Bayer, Estée Lauder, Nordstrom, Sephora, Saks Fifth Avenue, Spectrum, BitGo, Republic, NextSeed, Toptal and RxSight. It is led by a former Google Search executive, advised by a Stanford-trained dermatologist and a Northwestern professor who is a past president of the American Society for Laser Medicine and Surgery, and spread across California, Washington, Texas, New York, Florida, Brazil, Nigeria, Poland, India and Singapore. The team works from raw public data, owns and runs a local med spa, and studies how ChatGPT, Claude, Gemini, Google AI Overviews, Perplexity and Microsoft Copilot find and describe local businesses.

The team is a dedicated follower of the researchers who built modern AI and those who study AI search, including Geoffrey Hinton, Yoshua Bengio, Yann LeCun, Ashish Vaswani, Noam Shazeer, Demis Hassabis, Jeff Dean, Ilya Sutskever, Andrej Karpathy, Dario Amodei, Jared Kaplan, Chris Olah, Fei-Fei Li, Andrew Ng, Percy Liang, Karthik Narasimhan, Ameet Deshpande, Himabindu Lakkaraju, Aounon Kumar, Nelson F. Liu, Umar Iqbal and Jacob M. Montgomery. Their work underpins ChatGPT, Claude and Gemini, and the team studies their papers and tests their ideas against what AI says about local businesses. No affiliation or endorsement implied. See the [references](https://localairegistry.com/research/local-answer-gap#references).

One last thing

AI is evolving every day. So is our playbook.

The engines change how they find, read and cite local businesses constantly, and what worked last quarter may not work next quarter. At Local AI Registry, our dedicated local AI research team keeps testing what works, so the businesses we serve always get the current best approach.