When I talk to small business owners about AI search, I usually hear one of two beliefs. The first is hopeful: “AI is a fresh start, so it will treat my small shop the same as the big chains.” The second is gloomy: “AI just repeats what the big brands say, so there is no point trying.”
Both are half right. There is now real data comparing how strongly Google and the AI assistants lean toward well-known brands, and the gap between them is bigger than most people expect. But the same data also shows exactly where small brands still lose, and why.
Let me explain what the research found, why Google behaves differently from ChatGPT and Perplexity, and what a smaller business can do about it this quarter. I will also show you a simple way to measure whether bigger rivals are crowding you out of AI answers.
Do AI assistants really favour big brands less than Google?
In brief: the early evidence says yes, Google leans harder on brand size than ChatGPT and Perplexity do. But “less biased” does not mean “unbiased,” and it certainly does not mean small brands win by default.
The best direct comparison comes from Patrick Stox at Ahrefs. In Google Seems More Biased Towards Big Brands Than ChatGPT and Perplexity (published July 7, 2025), he asked a simple question: if a website is mentioned on more web pages, does it also show up more often in AI answers?
He looked at the top 50 sites in Ahrefs Brand Radar for June 2025 across three platforms. The dataset was large: about 76.7 million Google AI Overviews, 957,000 ChatGPT prompts and 953,500 Perplexity prompts.
Here is what he found, in plain words:
| Platform | Correlation between web mentions and AI visibility | What it means in plain words |
|---|---|---|
| Google AI Overviews | 0.65 | Strong. Sites that are talked about more across the web tend to appear much more often. |
| Perplexity | 0.30 | Weak. Fame helps a little, but plenty of less-mentioned sites get in. |
| ChatGPT | 0.15 | Very weak. How famous a site is tells you very little about how often ChatGPT mentions it. |
Source: Ahrefs, July 2025, using Spearman rank correlation on the top 50 sites per platform.
“Google has a strong correlation, where it seems like branded web mentions lead to more visibility.”
— Patrick Stox, Ahrefs, Google Seems More Biased Towards Big Brands Than ChatGPT and Perplexity
How to read these numbers without overreaching
A correlation runs from 0 (no relationship) to 1 (a perfect match). A score of 0.65 means two things rise together quite reliably. A score of 0.15 means they barely move together at all.
I want to be careful here, and so was Ahrefs. This is correlation, not proof. The study used only the top 50 sites per platform, which are all big websites anyway. It tells us that, among the most visible sites, Google's choices track fame much more closely. It does not prove that ChatGPT will pick your ten-person firm over a national chain. I treat it as direction, not hard science.
The finding also fits a larger Ahrefs study of 75,000 brands. In An Analysis of AI Overview Brand Visibility Factors (May 26, 2025, updated April 2026), branded web mentions had the strongest link with AI Overview visibility at 0.664. Brands in the top quarter for mentions averaged 169 AI Overview mentions, while the bottom two quarters averaged between 0 and 3. If you want the full list of factors, I cover them in what drives brand visibility in AI answers.
Why Google has leaned toward well-known brands for years
None of this is new behaviour for Google. Its AI Overviews are built on top of its normal search systems, so they inherit the habits of those systems.
Google's own documentation is clear on this. Its page AI features and your website (updated December 2025) says a page only needs to be indexed and eligible to show with a snippet to be used as a supporting link, and that “there are no additional requirements to appear in AI Overviews or AI Mode.” In other words, the same ranking systems that pick the ten blue links also feed the AI answer.
And those systems are built around trust. Google's guide to helpful content puts it in one line:
“Of these aspects, trust is most important.”
— Google Search Central, Creating helpful, reliable, people-first content
The same page asks site owners whether their site is “well-trusted or widely-recognized as an authority on its topic.” For a search engine serving billions of queries, being widely recognised is a cheap and safe proxy for trust. A brand that thousands of other pages talk about is less likely to be a scam.
Ahrefs' article also points back to former Google CEO Eric Schmidt's well-known view that brands are part of the answer to bad information online. Whatever you think of that, it explains the pattern: Google has spent many years building systems that reward being known.
So when a small business says “Google only shows the big guys,” there is something to it, especially on broad, high-value searches. I have written about where Google still leaves gaps in my guide to finding weak spots in Google's results.
Why ChatGPT and Perplexity can surface a smaller business
The assistants work differently. When ChatGPT or Perplexity searches the web, it usually turns your question into one or more shorter searches, reads the pages it finds and then writes an answer. The final pick depends heavily on which page answers the specific question best, not only on which brand is biggest.
Three pieces of evidence support this.
- Assistants cite deep, specific pages. Similarweb's April 2026 study of U.S. citations found that AI rarely cites homepages. Most ChatGPT citations came from pages several folders deep, such as guides, FAQs and product pages.
- Smaller sites get a bigger slice of their traffic from AI. In Ahrefs' March 2025 study of about 35,000 sites, sites under 10,000 visits got about 0.3% of their traffic from AI tools, compared with 0.1% for sites over 1 million visits.
- The pattern held in an earlier sample too. Ahrefs' February 2025 study of 3,000 sites found the smallest group (under 999 monthly visitors) got 0.37% of views from AI, against 0.09% for sites over 10,000 visitors.
“AI favors detailed, specific pages over broad landing pages.”
— Similarweb, The Most Cited Domains by LLMs and the Signals That Get You Cited
Please read the traffic numbers correctly. Big sites still get more AI visits in total. The point is that, as a share of their traffic, small sites are not shut out. A focused page from a small business can earn a citation that a big brand's generic page does not.
Why each platform behaves a little differently
Ahrefs describes Perplexity as running a real-time web search, then ranking, quoting and citing the best match, so its sources follow the results it pulls in. Ahrefs also noted that Perplexity's top sources include many local and regional brands. ChatGPT mixes what the model learned in training with live search, and it leans on Wikipedia and news publishers. I go deeper into these differences in why each AI assistant cites different sources.
Where the long tail gives small brands a real opening
The opening for a small brand is not on the broadest questions. It is on the specific ones. Here is how I think about it.
| Type of question | Example | Who usually wins | Small-brand chance |
|---|---|---|---|
| Broad category | “Best CRM software” | Big brands and big review sites | Low |
| Narrow use case | “Best CRM for a 3-person roofing business” | Pages written for that exact use | Medium to high |
| Local | “Emergency boiler repair in Leeds open Sunday” | Local firms with clear pages, reviews and listings | High |
| Expert how-to | “How to stop condensation on single-glazed sash windows” | The clearest, most specific guide | High |
| Brand vs brand | “Brand A vs Brand B” | Comparison sites, Reddit, the brands themselves | Medium |
The table is my own framework, not a study result. But it follows from the evidence above: when the question is narrow, there are fewer pages that answer it well, and size matters less.
Good long-tail targets usually have these features:
- A specific customer type (“for dentists,” “for first-time landlords”).
- A specific place, down to the town or neighbourhood.
- A specific problem or constraint (budget, size, deadline, material).
- A question big brands find commercially boring to answer in depth.
Where small brands still lose in AI answers
Now the honest part. “Weaker correlation with fame” does not mean fame stopped mattering. Several signals still favour big players heavily.
Links and traffic still line up with ChatGPT citations. An SE Ranking study reported by Search Engine Journal (November 26, 2025) looked at 129,000 domains. Sites with up to 2,500 referring domains averaged 1.6 to 1.8 ChatGPT citations, while sites with over 350,000 averaged 8.4. Sites under 190,000 monthly visitors averaged 2 to 2.9 citations, while sites over 10 million averaged 8.5.
Review platforms matter. The same study found domains listed on multiple review platforms averaged 4.6 to 6.3 citations, against 1.8 for those missing from them. A small business with no review footprint has little for an assistant to quote. I explain this in how review data shapes AI recommendations.
The most-cited sources are giant platforms. Ahrefs' June 2025 data shows Wikipedia leading ChatGPT at 16.3% of mentions and YouTube leading Perplexity at 16.1%. Semrush's 2025 study found Reddit and LinkedIn among the top five cited domains on ChatGPT, AI Mode and Perplexity. These are places where big brands are discussed far more often.
“Best of” lists favour the familiar. When an assistant builds a shortlist, it often borrows it from comparison articles. If those articles only list the big names, so will the answer. Getting onto those lists is a job in itself, which I cover in how to get your brand into “best of” lists.
And if your brand is brand new, the challenge is different again. The model may not know you exist at all. My guide on AI visibility for new brands deals with that specific problem.
A worked example: a small HVAC firm against a national chain
Let me make this concrete. This is a hypothetical example, with made-up numbers, to show how the maths works.
Imagine a 12-person heating and cooling company in one city. A national chain also serves that city. The owner writes 40 questions customers really ask, and runs each one in ChatGPT, Perplexity and Google AI Overviews. That is 120 answers in total.
| Question group | Answers checked | Chain named | Local firm named |
|---|---|---|---|
| Broad (“best AC brand,” “how much is a new furnace”) | 30 | 18 | 1 |
| Local (“furnace repair in [city] same day”) | 45 | 20 | 14 |
| Specific how-to (“why does my heat pump ice up in [city] winters”) | 45 | 6 | 9 |
| Total | 120 | 44 (37%) | 24 (20%) |
Overall, the chain looks dominant. But look by group. On broad questions, the local firm is almost invisible (1 in 30). On specific how-to questions, it is already ahead (9 vs 6). On local questions it is behind, but within reach.
So the plan writes itself. Do not fight for “best AC brand.” Build more specific guides, fix local listings and reviews to close the local gap, and measure again in a month. That is how a small business uses the weaker brand bias in practice.
Seven moves for a small business competing with bigger names
Here is how I would do it, in order.
- List your winnable questions. Pull 30 to 50 questions from sales calls, emails and Search Console. Mark each as broad, local or specific.
- Build one strong page per specific question cluster. Answer the question in the first paragraph, then give detail: prices, steps, local factors, photos. Specific pages are what assistants cite.
- Fix the basics that every platform checks. Make sure your pages are indexed in Google and Bing, and that you are not blocking AI search crawlers by mistake.
- Grow your review footprint. Ask every happy customer for an honest review on the two or three platforms your buyers use. Never buy or fake reviews.
- Earn mentions in local and trade media. A quote in a city news site or trade magazine is the kind of web mention that helps on Google, where fame matters most. My guide to digital PR for AI search explains how.
- Get onto the lists that already rank. Find the “best [service] in [city]” articles that assistants cite for your questions, and pitch the authors with real reasons to include you.
- Measure every month. Re-run the same questions and track your share. Small brands win by steady gains on specific questions, not one big jump.
If you serve a local area, most of steps 3, 4 and 6 overlap with good local SEO work anyway.
Measuring your share of AI answers against bigger rivals
You cannot improve what you do not measure. The simplest metric is share of AI answers:
Share of answers = answers that name your brand ÷ total answers checked × 100
Track the same number for two or three rivals, including the biggest one. Then split it by platform and by question type, as in the example above. The split tells you far more than the total.
| What to track | How to collect it | Why it matters |
|---|---|---|
| Share of answers per platform | Manual checks or an AI visibility tool | Shows where the brand bias hits you hardest |
| Share by question type | Tag each question broad, local or specific | Shows where you can realistically win |
| Citations to your own pages | Note which URL is linked, if any | Tells you which pages are working |
| Third-party sources cited | Record the domains cited next to rivals | Gives you an outreach list |
| AI referral visits | Analytics (GA4 now has an AI Assistant channel) | Connects visibility to real visitors |
For a deeper method, see my guide to measuring AI share of voice, and my comparison of AI visibility tools if you would rather not do it by hand. To judge whether your AI visits are normal for a site your size, use my AI referral traffic benchmarks.
One tip: answers change from run to run. Run each question at least twice, or in a fresh session, and count a brand as “named” only if it appears in most runs.
Questions small business owners ask about AI and big brands
Is ChatGPT fairer to small businesses than Google?
In one narrow sense, the data points that way. Ahrefs found that how often a site is mentioned across the web lines up strongly with visibility in Google AI Overviews (0.65) but only very weakly with ChatGPT (0.15). That is not the same as fair. Small brands still need clear pages, reviews and third-party mentions to be picked.
Does a big brand automatically win in AI answers?
No. Big brands start with more mentions, links and reviews, so they win many broad questions. But for narrow, specific or local questions, assistants often cite the page that answers the question best, and that can be a small specialist.
Should a small business focus on AI search instead of Google?
No. Google still sends far more traffic to most sites, and AI Overviews and AI Mode draw from Google's own index. Treat AI assistants as an extra place to be found, built on the same foundations as your SEO.
How can I tell if bigger competitors are crowding me out of AI answers?
Write down 30 to 50 questions your customers ask, run them in ChatGPT, Perplexity and Google, and record which brands are named and cited. Repeat each month. Your share of those answers, compared with your rivals', is the number to watch.
Do I need a Wikipedia page to compete with big brands in AI answers?
Not necessarily. Wikipedia is heavily cited, but most small businesses do not meet its notability rules. Reviews, accurate directory listings, trade press, community threads and well-built deep pages on your own site are more realistic routes.
Where to start if you are the smaller brand in your market
So, to bring it together. Google's AI answers follow brand fame closely, because they sit on top of search systems built to reward trust and recognition. ChatGPT and Perplexity follow fame much less closely, at least in the data we have so far. That gives smaller businesses a genuine opening, but mainly on specific, local and expert questions, and only if you have the pages, reviews and mentions to back it up.
My suggestion for this week: write down your 30 most common customer questions, run them through the three platforms, and mark where you already appear. That one exercise will show you where the big brands are beatable.
If you would like help, our team can run that audit and build the plan for you. Take a look at our SEO services, or get a free quote and tell us who you are up against.
Further reading on big brand bias AI search
- Why Each AI Assistant Cites Different Sources
- AI Visibility Tools Compared: Options for Tracking Brand Mentions in AI
- AI Referral Traffic Benchmarks: How Fast It's Growing and What's Normal
Sources for this big-brand bias in AI search guide
- Ahrefs: Google Seems More Biased Towards Big Brands Than ChatGPT and Perplexity (July 7, 2025)
- Ahrefs: An Analysis of AI Overview Brand Visibility Factors, 75K brands (May 26, 2025, updated April 2026)
- Ahrefs: AI traffic research across ~35,000 sites (March 26, 2025)
- Ahrefs: AI traffic study across 3,000 sites (February 6, 2025)
- Ahrefs: The 10 most mentioned domains for ChatGPT, Perplexity and AI Overviews (June 11, 2025)
- Google Search Central: Creating helpful, reliable, people-first content (updated October 5, 2026)
- Google Search Central: AI features and your website (updated December 10, 2025)
- Similarweb: The most cited domains by LLMs (April 14, 2026)
- Search Engine Journal: New data on top factors influencing ChatGPT citations (SE Ranking study, November 26, 2025)
- Semrush: The most-cited domains in AI, a 3-month study (November 10, 2025)
