How to Track Your Local Business in ChatGPT: A Step-by-Step Setup for 2026

Key takeaways

• Start with a manual baseline: 15 to 25 realistic local prompts, run weekly, logged in a spreadsheet. It costs nothing and teaches you more than any dashboard on day one.
• Track three kinds of prompts separately: category searches ("best emergency dentist in Seattle"), comparison prompts, and branded prompts ("is [your business] any good?").
• ChatGPT and Gemini often disagree on local recommendations. Searchable's data found they agreed on only about a quarter to a third of local service businesses, so check each engine on its own.
• ChatGPT now searches specific domains directly, so your own site needs to be crawlable. Platform behavior changes overnight, which makes monitoring better than a one-off audit.
• Once you have more than a handful of prompts, move to an automated tracker. Several in the catalog below handle this.

Why tracking local visibility in ChatGPT is different

A customer types "which dentist in downtown Seattle takes emergency appointments?" into ChatGPT. It answers with two or three names, a few details, and sometimes a short summary of reviews. There is no page of ten blue links and no local 3-pack you can screenshot. Either your business is in the answer or it isn't.

That makes tracking awkward. Answers vary between runs, between users, and between locations. A rank tracker can't tell you much here. What you can measure is how often you appear across a stable set of prompts, how you are described, and which sources ChatGPT leans on to describe you.

One caveat before we start. The data I cite from Promptwatch below comes mostly from software and SaaS prompts, not local ones. I'll flag where that matters. Treat it as a guide to how ChatGPT builds answers in general, not as local benchmarks.

Step 1: Define what you are tracking

Before opening ChatGPT, write down four things.

• Your exact business name and any common variations or misspellings.
• Your service area: city, neighborhoods, and the nearby towns customers actually come from.
• Your two or three core services. Be specific. "Emergency pipe repair" beats "plumbing."
• Your three to five closest local competitors.

This list becomes the scoring sheet. A mention only counts if it is accurate, so you also need the correct address, phone number, hours and services written down to compare against.

Step 2: Build a local prompt list

Aim for 15 to 25 prompts. Fewer than that and one odd answer skews everything. More than that and the weekly run becomes a chore you'll skip.

Split them into three groups.

Category prompts

These are the ones that win new customers. Customers don't know you yet.

• "Best [service] in [city]"
• "Who is the best [service] near [neighborhood] for [specific need]?"
• "[Service] in [city] open on weekends"
• "Affordable [service] in [city] with good reviews"

Vary the phrasing. People type short, sloppy queries into ChatGPT and long conversational ones, and the results differ.

Comparison prompts

• "[Your business] vs [competitor]"
• "Is [competitor] better than [your business] for [service]?"
• "Alternatives to [competitor] in [city]"

Branded prompts

These test whether ChatGPT knows you at all.

• "What is [your business] in [city]?"
• "Is [your business] good?"
• "What are [your business]'s hours and services?"
• "What do customers say about [your business]?"

Branded prompts matter more than they look. In Promptwatch's ChatGPT content type data (Sep 28 to Oct 4, 2026), reviews made up 5.6% of citations on branded prompts, more than three times their overall 1.7% share. Product pages (23.3%) and landing pages (22.8%) led. The examples in that report are SaaS, but the pattern is plausible for local: when someone names your business, ChatGPT goes looking for your own pages and for reviews.

Step 3: Run the manual baseline

Open a fresh ChatGPT session, ideally logged out or in a temporary chat so your own history doesn't color the result. Run each prompt and record what happens.

A spreadsheet is enough. If you want something a bit more structured, a free Airtable base works, and Zapier or Make can push results into it later.

Run each prompt twice if you can. ChatGPT is not deterministic, and a single run can mislead you. If you appear in one of two runs, call it a partial win, not a hit.

The two columns that pay off most are competitors and sources. Competitors tell you who you are really up against in AI answers, which is often different from who you compete with on Google Maps. Sources tell you where to work. If the same directory, local news site or review platform keeps getting cited, that is where your listing needs to be accurate and current.

Step 4: Compare engines, not just ChatGPT

Do the same run in Gemini and, if you have time, Perplexity. Local businesses really do get different treatment across engines. Searchable's research found ChatGPT and Gemini agreed on only about a quarter to a third of the local service businesses they recommended for the same questions, versus close to half for B2B software brands.

That has a practical consequence. If you show up in Gemini because of a strong Google Business Profile but never in ChatGPT, the fix is probably not the same as if you were missing from both. Gemini leans on Google's own data. ChatGPT leans on the open web, including Bing-style search results, your site and third-party pages.

For broader context on how engines differ, Promptwatch's social media citations by AI model report shows ChatGPT drew just 0.46% of its citations from social platforms between Sep 28 and Oct 4, 2026, while Google AI Overviews drew 12.2%. Same question, very different source mix.

Step 5: Check that ChatGPT can actually read your site

Tracking tells you the symptom. This step checks one common cause.

On August 8, 2026, ChatGPT Search began using the site: operator at scale. Promptwatch Data shows site:-scoped fanout queries jumping from 0.37% to 16.8% of all fanouts overnight, roughly one in six, and average searches per response rose from about 1.08 to about 1.83. The site: searches were added on top of generic searches rather than replacing them.

What that means for a local business: ChatGPT will now often search your domain directly while answering. Your site is a retrieval surface. If it is blocked, thin, slow, or mostly images of a menu in a PDF, you give it nothing to work with.

A short checklist:

• Robots.txt doesn't block OAI-SearchBot or ChatGPT-User. Both matter for search-style answers.
• Every service has its own page with a plain-language description, the area served, and prices or price ranges where you can state them.
• Name, address, phone and hours appear as text, not only in an image, and match your Google and Bing listings exactly.
• There is a short FAQ page answering what customers actually ask on the phone.
• Pages are indexed in Google and Bing. A technical crawler such as Screaming Frog or Sitebulb will find broken pages and orphans.

Promptwatch's query fanout data also suggests ChatGPT's generated searches have gotten shorter, with average query length falling from about 117 characters in December 2025 to roughly 53 in April 2026. Their advice: write headings that read like short search queries and put the entity and category first. For a local business that means a heading like "Emergency plumber in Leeds" beats "Our story."

One more point from that data: fanout keyword mappings built before August 8 are stale. If you built a prompt list last spring from some guide, rebuild it.

Step 6: Fix the sources ChatGPT trusts about you

The manual run gives you a list of cited sources. Work through them.

1. Match your Google Business Profile and Bing Places listings to your website, character for character.
2. Claim or correct the three or four directories that keep appearing in the cited sources.
3. Ask for fresh reviews on the platforms ChatGPT cites. Steady and genuine beats a burst.
4. Get mentioned on local sites: the chamber of commerce, local news, community blogs, partner businesses.
5. Add LocalBusiness and Service schema in JSON-LD. Google says no special markup is required for its own AI features, so treat schema as helpful clarity, not a magic switch.

Listing consistency tools like Yext help if you manage many locations. SOCi covers local search, social and reviews for multi-location brands. For a single shop, doing it by hand takes an afternoon.

Birdeye is worth a look if reviews are your weakest area.

Step 7: Decide when to automate

The spreadsheet method works for 20 prompts across one location. It breaks when you add a second location, want daily data, or want to see more than a pass or fail. Here is how the options in our catalog line up for a local business.

Promptwatch is the one I would look at first if you want more than a pass or fail. Its tracking uses the real ChatGPT interface rather than API output, which matters because user-facing answers and citations can differ from API responses. City and state tracking starts on the Business plan ($579/mo), which is the part that makes location-specific answers testable. Country tracking is on every paid plan, and the Essential plan starts at $95/mo with a 7-day free trial on monthly billing. Confirm current numbers at promptwatch.com/pricing.

The reason to pay for something like this is the layer after tracking. IP-verified crawler logs show whether ChatGPT's bots actually reach your pages and where they hit errors, and the Action Board lists what to fix. For a local business that is overkill at first, so start with the free Explore plan or the free AI Visibility Report to see whether it clicks.

Cheaper options exist and are fine for a single location.

If you already pay for Semrush, check what its AI tracking covers before adding another subscription.

For a wider look at what is available, the GEO software directory at bestgeosoftware.com lists far more tools than fit here.

Step 8: Read the results correctly

A few rules that save you from bad conclusions.

• Look at the trend over four weeks, not a single week. Local answers wobble.
• A mention without a citation to your own site usually means ChatGPT learned about you from elsewhere. That is a sign to improve your own pages.
• A citation without a mention means ChatGPT read your page but didn't recommend you. Check whether the page actually answers the question.
• Competitor wins are information. Open the pages ChatGPT cited for them and see what those pages state that yours doesn't.
• Wrong facts are the urgent problem. If ChatGPT gives the wrong hours or a closed location, find the source and fix it there.

Also keep an eye on how long content stays cited. Promptwatch's content half-life data tracked pages ChatGPT cited on July 6, 2026 for 90 days. Landing pages and docs stayed cited longest, while news lost half its pages within 36 days. For a local business, that argues for solid evergreen service pages over a stream of news-style posts.

And don't expect uniform sources. Promptwatch's unique domains report found almost half of ChatGPT responses cite 10 or more unique domains (Sep 28 to Oct 4, 2026). That is a wide net, and it means being named on several trusted pages helps more than perfecting one.

Step 9: Set a routine

Here is a cadence that a busy owner can keep.

The last row is real. ChatGPT changed its search behavior overnight in August 2026. Promptwatch's citation drop report shows average citations per response falling across GPT-5.3, GPT-5.4 and GPT-5-Mini around the March 4, 2026 rollout. A quiet week of tracking can follow a big shift, so don't assume last quarter's results still hold.

Common mistakes

• Testing only your brand name. Of course ChatGPT knows you when you type your own name. Category prompts are where customers are won.
• Using one phrasing per query. Test three versions of your main category prompt.
• Running prompts while logged in to a heavily personalized account.
• Ignoring wrong information because you are still being mentioned.
• Treating schema as a cure. It helps clarity. It does not replace reviews, listings and a site worth citing.
• Paying for an enterprise platform to track 15 prompts for one shop.

When to get help

If you manage several locations or compete in a crowded category like legal or dental, the work of fixing listings, building local mentions and publishing service content adds up. Agencies such as 1001 SEO Media handle GEO and local SEO together, which saves you coordinating two vendors. If you would rather do it yourself, the steps above are enough to get started this week.

A starting point for this week

Write 20 prompts today, run them tomorrow, and put the results in a sheet. Fix any wrong facts you find before you do anything else. Then decide in a month whether the weekly run is worth automating.

For more reading on the approach, eSEOspace's guide on making ChatGPT recommend your local business is a reasonable companion for the schema and entity side of things.