How Do Franchise Brands Get Recommended by ChatGPT?

ChatGPT recommends franchise locations, not franchise brands, and the numbers show how little a national name carries on its own. In the Cheers Market Baseline, national franchise brands appeared in 23.8% of the AI answers that named a home-services business, but they filled 7.5% of the named slots and were named first in 8.3% of answers. On ChatGPT the figures were lower still: 21.1% of answers, 5.0% of slots, first in 4.6%. Every week we ask ChatGPT, Gemini, Perplexity and Google AI Mode the same 32 home-services questions in 100 US and Canadian metros, and these numbers cover 34,742 answers from August 31 to September 20, 2026. For a franchise marketer, that means the brand name shows up in roughly one answer in four, and after that each location competes on its own profile, reviews and page, one market at a time.

Last verified: September 23, 2026.

The International Franchise Association expects franchise establishments in the US to grow from 832,521 to 845,000 in 2026. Each one is a separate hiring decision for a customer and a separate lookup for an answer engine. The Cheers franchise SEO program is built for that unit of work: the individual location, measured and fixed one at a time.

How often national brands show up depends on the trade

National brands were nearly absent from roofing answers (0.9%) and rare in HVAC (9.3%), but they appeared in 69.4% of water damage restoration answers and 60.1% of pest control answers. In between sat windows and doors at 47.6%, garage door at 40.6%, plumbing at 36.4%, duct cleaning at 29.4%, electrical at 13.3% and lawn care at 10.3%. Six of those trades rest on a single question each, so read their numbers as answers to that question rather than the whole trade. US answers included a national brand more often than Canadian ones, 27.8% against 17.4%.

We count brands as a category and never name one; the brand list stays internal. The shares also lean high, for two reasons: some answers mention national chains only to tell readers to avoid them, and franchise names are easier for our extraction to catch than a small independent's.

For a franchise marketer the spread sets expectations by concept. In restoration, a national brand in the answer is normal and the question is which location gets named. In roofing or HVAC, a national name is rarely in the answer at all, and each location competes with independents on local evidence. The trade pages go deeper, starting with how restoration companies get recommended by ChatGPT, the trade where brands show up most. Even inside one trade the share swings by job: in plumbing it ran from about half of the answers on urgent drain and burst-pipe questions to under a sixth on tankless water heater installation, as how plumbing companies get recommended by ChatGPT shows question by question.

Showing up is not the same as being picked

Even where brands appear, they fill a small part of the answer. Across all 32 questions, national franchise brands filled 7.5% of the 134,718 named business slots in 33,341 answers that named at least one business. Independent and regional companies filled the other 92.5%.

The engines differ in degree, not direction. Perplexity gave national brands 10.0% of its named slots, Gemini 8.9%, Google AI Mode 8.4% and ChatGPT 5.0%. Google AI Mode was the most likely to include a brand at all (27.2% of its answers), and Perplexity the most likely to name one first (13.0%), while ChatGPT sat at the bottom on all three measures: 21.1% of answers, 5.0% of slots and first in 4.6%. Gemini included a brand in 22.9% of its answers and named one first in 9.5%; Perplexity included one in 24.0%; Google AI Mode named one first in 7.6%.

Treat the Perplexity figures as indicative. Its extraction recall was the lowest of the four engines and a collector outage cut its answer count.

AI engines name the location, so compete location by location

Everything that decides whether a location gets named sits at the location. Google's own guide to its AI features says "Using products like Merchant Center and Google Business Profiles can help your products and services to be visible in both AI responses and other Google Search results." In our data, Google AI Mode showed rated business cards in 90% of its answers, and each card is a single location's profile, with that location's rating and review count. Order follows review count: in Google AI Mode answers that named three or more rated businesses, the first one had a median of 342 reviews against 195 for the third, and the most-reviewed business came first 36% of the time against 25% by chance. That is a correlation, since review count also tracks size and age, but the reviews that count are the location's, not the brand's.

The pages work the same way. Of 74,969 citations to home-services business websites, location or city pages took 14.0% and service-plus-city pages 17.1%, 31.1% between them, while blog and resource pages took 1.5%. On Perplexity, service-plus-city pages were the single most-cited page type at 28.6%. Google AI Mode sent 66.9% of its business-site citations to homepages, probably through the website button on each business profile, though that is our inference rather than something the data shows. Google's own guideline for chains points the same way: "provide a website that represents your individual business location." A location profile whose website button opens the brand homepage hands the answer a page with nothing about that market.

The same goes for the other listings. About a third of the rated business cards on ChatGPT's map (34%) linked to Yelp, so each location's Yelp listing is part of its public record too. Google AI Mode also linked an online booking page in 31% of its answers, and we think most of those links come from the booking link on each business's profile, which the data cannot confirm. And the engines rarely agree: on occasions when all four answered the same question for the same city in the same week, 77% of the businesses named were named by only one engine.

Google's rules for chains make the unit explicit. "Do not create more than one page for each location of your business," and "All locations must have the same name unless the business's real world representation consistently varies from location to location." The brand sets the standard; each location has to meet it in public.

Brand-level work still has a job

None of this makes the brand irrelevant. The brand site sets the naming standard, the canonical location URLs, the structured data template and the review program every location runs. What it cannot do is stand in for a location. Yext found the same thing from the other direction in its October 2025 research on AI citations and user location:

A retail chain might report a 47% first-party citation rate nationally, but location analysis could reveal rates of 70% in rural markets and 20% in competitive urban markets.

A national 47% describes no actual store. The customer is asking which location near them to trust, and so is the engine. For a rollup that bought local companies, the same logic applies to every legacy name, phone and profile still live in each market. The multi-location SEO program and multi-location local SEO cover the listing and page cleanup that goes with it.

Reviews scale the same way. Hello Sugar, a Cheers customer, runs one review program across its salon network; its September 7, 2026 snapshot covered 276 Google profiles and 70,581 Google reviews, with 34,860 of them dated since it joined Cheers. The program is national. The reviews land on each salon's own profile, which is where an answer engine reads them.

What to do this week

Pick the ten locations with the most revenue at stake, across your two largest service lines. For each one, run the two service questions a local customer is most likely to ask, such as "who are the best plumbers for emergency plumbing service?", on ChatGPT and Google AI Mode with the location's city added, and record which businesses were named and which sources were cited. Then check the location's own record against what you found:

• The Google Business Profile: one profile, the approved name, the right category, and a website button that opens that location's page rather than the brand homepage.
• The review count next to the named competitors in that market, because in these answers the first-named business tended to have more reviews than the rest.
• The location page: services the location actually performs, its service area, hours, phone and proof from that market.
• The booking link: it should open that location's scheduler.

Fix the gaps location by location and rerun the same questions a week later. Change nothing about the questions, engines or cities between runs, or the comparison stops meaning anything.

Methodology

Every week, Cheers asks ChatGPT, Gemini, Perplexity and Google AI Mode the same 32 home-services questions in 100 metro areas across the United States and Canada, and records every business each engine names, with its star rating, review count and map position where the engine shows one, and every source it cites. This analysis covers the three weeks from August 31 to September 20, 2026: 34,742 usable answers (9,434 from ChatGPT, 9,435 from Google AI Mode, 9,364 from Gemini and 6,509 from Perplexity) across 60 US and 40 Canadian metros. We excluded empty or cut-off answers, most of them from a Perplexity collection outage. Each question was asked once per city per week, so week-to-week changes include normal answer-to-answer variation. Business names were extracted automatically; a hand check of 40 answers found the extraction caught about three in four of the businesses a reader would count, so counts are lower bounds. Star ratings and review counts are the figures each engine displayed; for ChatGPT and Gemini we captured them from September 18 and September 14 onward. Findings describe associations in this sample, not how any engine ranks businesses.

The 32 questions cover plumbing (9), HVAC (9), electrical (6), roofing (2), and one question each for pest control, duct cleaning, garage door, water damage restoration, lawn care, and windows and doors. National franchise brands are counted as a category from an internal list of multi-brand franchise groups and national service chains, big-box installers and Canadian national comfort brands; no brand is ever published with a number. A slot is one named business in one answer. Earlier versions of this page quoted Cheers customer-panel rates; this version uses only the Market Baseline, which is Cheers-owned and published by name.

Sources

Every external link below was opened and checked on September 23, 2026.

• Cheers Research methodology. Cheers Market Baseline, August 31 to September 20, 2026, 34,742 answers, aggregate only. Source of every national-brand, engine, review, page-type, booking and agreement figure on this page.
• 2026 Franchising Economic Outlook. International Franchise Association, 2026. Source of the 832,521 and 845,000 establishment figures.
• Guidelines for representing your business on Google. Google Business Profile Help, undated. Source of the three quoted rules for chains and locations.
• AI Citations, User Locations, and Query Context. Yext Research, October 9, 2025. Source of the quoted 47%, 70% and 20% example.
• Optimizing your website for generative AI features on Google Search. Google Search Central, last updated July 10, 2026. Says Google Business Profiles can help products and services show up in AI responses, and that no special markup is required.
• Hello Sugar case study. Cheers, September 7, 2026 snapshot. Source of the 276 profiles, 70,581 reviews and 34,860 reviews since joining.

Dylan Allen-Arnegård is the CEO & Co-Founder of Cheers, the local search platform for multi-location service businesses.