How AI assistants decide which running store to recommend

A runner new to town types “where can I find a store that specializes in running gear” into ChatGPT, or asks Google for the best place to get a gait analysis. The answer names two or three stores. Here is how those names get picked, and what you control.

Where do AI assistants get local store information?

Every major assistant searches at the moment someone asks. None of them documents a list of “recommended running stores” you can apply or pay to join. They pull from search results, map data, reviews and the pages they can crawl, then write an answer and cite some of the sources.

Assistant	Where it gets local info	What you control
ChatGPT search	Its own crawler (OAI-SearchBot), plus search partners it sends rewritten queries to, plus third-party providers for nearby business listings	Allowing OAI-SearchBot, your website text, accurate listings on the sites that list you
Perplexity	A live web search with numbered citations, crawled by PerplexityBot	Allowing PerplexityBot, your website text, third-party pages that mention you
Gemini	Public Google Maps information (addresses, hours, ratings, reviews)	Your Google Business Profile, your reviews, your website
Google AI Overviews and AI Mode	Google’s search index, with no special requirements beyond normal Search eligibility	Indexable pages, structured data that matches the page, Business Profile and Merchant Center data

The details below come from each company’s own documentation.

How does ChatGPT find local businesses?

OpenAI’s help article on searching the web with ChatGPT explains the flow:

1. ChatGPT estimates the person’s general location from their IP address. Sharing a precise device location is optional and off by default.
2. When it uses a search partner, it typically rewrites the question into search queries for its search partners. OpenAI’s own example turns “good restaurants near me” into “top restaurants San Francisco.” The same article points readers to Microsoft’s and Shopify’s privacy statements for how those providers handle queries.
3. It may share that location with third-party providers that supply nearby business listings.
4. On iOS and Android, the answer may include a map.

The same article says ChatGPT ranks results using multiple factors and that placement is not guaranteed. To be eligible at all, OpenAI says to allow OAI-SearchBot to crawl your site and make sure your host or CDN does not block OpenAI’s crawler.

What this means for a run store: ChatGPT is searching for something like “running shoe store [your town]” or “gait analysis [your town].” Stores that show up well for those searches, on the web and in business listings, are the ones in the running to be named.

How does Perplexity find stores?

Perplexity says it searches the internet in real time and that each answer includes numbered citations linking to the original sources. Its crawler, PerplexityBot, exists to surface and link websites in Perplexity’s results, and Perplexity says it is not used to collect training data for AI models.

Perplexity’s documentation doesn’t describe a separate process for local businesses. Its citations can come from any page its search turns up: your site, review sites, local news, race websites and “best running stores in [city]” articles. If those pages describe your fit process and your run club, Perplexity has something to quote.

How do Gemini and Google AI Overviews pick local businesses?

Google’s two AI products work differently, but both sit on top of data you already manage.

Gemini. Google’s help page on finding places in Gemini says Gemini uses public information from Google Maps for addresses, descriptions, websites, ratings and opening hours, and people can open a place to read reviews. For your store, that Maps information is what your Google Business Profile shows: hours, website, rating and reviews.

AI Overviews and AI Mode. Google’s page on AI features and your website is direct: to be a supporting link, a page must be indexed and eligible to show in Google Search with a snippet. Google says there are no additional requirements and no special optimizations needed. Both features may run several related searches across subtopics to build an answer, which Google calls query fan-out.

Google’s own list of what helps includes:

• Allowing crawling in robots.txt and in your CDN or hosting setup
• Making important content available as text
• Making sure structured data matches the visible text on the page
• Keeping your Merchant Center and Business Profile information up to date

That last point matters for run stores. Your Business Profile and your product data feed Google’s answers about who is open, nearby and carrying the shoe.

What do AI assistants read about your store?

Across all four, five sources matter.

Your Google Business Profile

Gemini reads Google Maps directly, and Google tells site owners to keep the Business Profile current for AI features. The details that matter most for a run store: the right primary category, hours that match reality (including race-day and holiday changes), services like gait analysis and shoe fitting listed by name, and photos of the shoe wall and the fit area. Our Business Profile checklist walks through every field.

Your reviews

Assistants summarize what reviewers say, and Gemini offers an AI summary of a place right next to its reviews. For example, a review that says “they watched me run, measured both feet and found a wide toe box that fixed my blister problem” gives an assistant a reason to name you for “best store for a gait analysis.” A wall of five-word reviews does not. BrightLocal’s 2026 consumer survey of 1,002 US adults found 45% said they used AI tools for local recommendations, up from 6% a year earlier. The figures are self-reported. You can’t script what customers write, but a memorable fitting tends to get described in detail. Our guide to getting more Google reviews covers how to ask within Google’s rules.

Your website’s text

Assistants quote text. If your fit process, your run club schedule, the brands on your wall and the neighborhoods you serve only live in images or a PDF, they are invisible. Google lists “important content available in textual form” as a basic practice for its AI features. Write a plain page that answers the questions runners actually ask: do you do gait analysis, does it cost anything, how long does a fitting take, do you carry trail and wide sizes, when is the group run.

Structured data

Structured data is code on your pages that labels facts like your address, hours, products and prices. Google says AI features need no special schema, but it does say your structured data should match what’s visible on the page. Accurate store and product markup gives machines a labeled copy of the same facts your Business Profile and product pages show. It won’t get you named on its own.

Third-party lists and mentions

“Best running stores in [city]” articles, race sponsor pages, run club pages, the local paper’s coverage of your shoe drive, and directory listings all feed the web results assistants search. Make sure your name, address and phone match everywhere. Mismatched details give an assistant two versions of you to choose from.

Are you blocking AI crawlers in robots.txt?

Your robots.txt file tells crawlers what they may read. On Shopify, every store gets a default robots.txt that you can view at yourstore.com/robots.txt. It is changed only through a robots.txt.liquid template, and Shopify warns that incorrect use can cause loss of all traffic. If someone edited yours, or a security service sits in front of your site, check it.

Know which crawler does what:

Crawler	Company	What it does	If you block it
OAI-SearchBot	OpenAI	Finds pages for ChatGPT search	Your site is not shown in ChatGPT search answers
GPTBot	OpenAI	Collects training data	Opts you out of training; search is unaffected
ChatGPT-User	OpenAI	Visits a page when a user asks	robots.txt may not apply
PerplexityBot	Perplexity	Finds pages for Perplexity answers	Perplexity recommends allowing it to appear in results
Perplexity-User	Perplexity	Visits a page when a user asks	Generally ignores robots.txt
Google-Extended	Google	Controls Gemini training and grounding use	Does not affect inclusion or ranking in Google Search

Sources: OpenAI’s crawler documentation, Perplexity’s crawler documentation, and Google’s common crawlers page. OpenAI and Perplexity both say robots.txt changes can take up to about 24 hours to reach their systems.

For a store that wants to be recommended, the safe default is simple: allow OAI-SearchBot, PerplexityBot and Googlebot. Whether you allow the training crawlers is a separate choice that does not change search visibility.

Do you need an llms.txt file?

No, not for being recommended as a store. llms.txt describes itself as a proposal: a Markdown file at your site’s root that gives AI agents a short summary of the site with links. It has not been adopted as a standard.

None of the documentation above from OpenAI, Perplexity or Google says they use llms.txt to choose sources. Google goes further on its AI features page: you don’t need to create new machine readable files, AI text files or markup to appear in AI Overviews or AI Mode. Adding one won’t hurt, but spend the hour on your Business Profile and your fit page first.

How do you test whether AI assistants recommend your store?

Run this once a month. It takes about 20 minutes.

1. Use a phone in your town. Assistants use your approximate location. Testing from the store’s Wi-Fi is fine; testing from a vacation is not. Use a private window or signed-out session where you can, so your own history and saved memories don’t skew the answer.
2. Ask the questions runners ask. Use the same five in every assistant:
   • “Where can I find a store that specializes in running gear near me?”
   • “Best running shoe store in [your town]”
   • “Where can I get a gait analysis in [your town]?”
   • “Where can I try on [a shoe you carry] today in [your town]?”
   • “Running store with a group run in [your town]”
3. Ask ChatGPT, Perplexity, Gemini and Google (look for the AI Overview or open AI Mode).
4. Record three things for each answer: which stores are named, in what order, and which sources are cited.
5. Read the sources. If a competitor is named and the cited source is their fit page or a “best of” list you’re missing from, that is your to-do list. If you’re named with the wrong hours or an old address, fix the source it cites.
6. Repeat in a month and compare.

OpenAI’s own tip for better local results is to include a city, neighborhood or postal code in the question, so test both “near me” and named-town versions.

What should you fix first?

Work in this order. Most of these steps help every assistant at once.

1. Business Profile. Correct primary category, real hours, services listed, fresh photos. This feeds Gemini and Google’s AI answers directly.
2. Reviews. Ask every customer for an honest review. Google doesn’t allow asking for specific content, so let the fitting do the talking.
3. A fit page in plain text. What a fitting includes, how long it takes, what it costs, and who it’s for.
4. robots.txt. Confirm OAI-SearchBot, PerplexityBot and Googlebot can read your site.
5. Listings everywhere else. Same name, address and phone on every directory, race page and club page.

If you want help with the first step and the code behind it, Get Found is the Segments setup that optimizes the Business Profile for every location, runs a full catalog pass and adds a structured data baseline to your Shopify site. To see where your site stands before you change anything, run the 1-minute test for your website.