• AI visibility

Guide · AI search

How can a small brand get recommended by AI assistants?

Mainly by being written about on independent sites and by stating a clear, checkable advantage, not by tweaking pages for each assistant. Small brands start far behind: in one vendor’s data they appeared in about one unbranded AI answer in ten. The research suggests building that wider presence first, then fine-tuning.

By Underneath Team Guide Updated

The short version

  1. In one vendor’s tracking data, small brands appeared in 9.9% of unbranded ChatGPT answers on day one, 8.9% on Gemini and 6.8% on Perplexity (Kumar (opens in a new tab), 2026).
  2. The same study’s fastest-rising small brands gained only 10 to 20 points between March and May 2026, so its author advises building broad presence before per-assistant tactics.
  3. In our brand entity study, the number of independent websites naming a brand was the strongest signal: 4.7 times the odds of a recommendation per tenfold increase.
  4. In controlled tests, an unknown brand beat a famous one half the time with a price just 7.3% lower (Chu and Hou (opens in a new tab), 2026).
  5. Asking in a brand’s home language raised how often AI assistants named local champions by 0.80 on a 0 to 1 scale, against 0.15 for multinationals (Żatuchin (opens in a new tab), 2026).

How far behind do small brands start?

Far behind, on every major assistant measured. Kumar (opens in a new tab), a co-founder of the AI visibility company Ranqo, grouped brands tracked on its platform from March to May 2026 into three tiers by stature. On each brand’s first tracking run, niche and small brands appeared in 11% of unbranded category answers.

Outside the two top tiers, the picture was similar everywhere: 9.9% of unbranded answers on ChatGPT, 8.9% on Gemini and 6.8% on Perplexity. Small brands are hard to surface on every assistant, not just on one.

Nor did they climb much on their own. The fastest-rising small brands moved only 10 to 20 percentage points across the observation window. This is a vendor’s customer data, with tiers assigned by hand, so treat the exact figures as indicative.

Why does being written about matter more than your own website?

Because AI assistants mostly build answers from other people’s pages. In Kumar’s data, “best of” lists were the most cited kind of content, at about 21% of all citations. A single list that includes your brand can be reused across many different questions. Among sources other than company websites, YouTube was cited most, at 4.2% of citations.

Our brand entity study found the same thing from another angle. Across ChatGPT, Gemini, Perplexity and Claude, the signal most closely tied to being recommended was independent coverage: how many websites, other than the brand’s own, named it in the pages the assistants cited. Each tenfold increase went with 4.7 times the odds of being recommended.

The other route in is genuine community discussion. In Sharma’s study (opens in a new tab) of 112 Product Hunt startups, links from other websites and real Reddit discussion went with being surfaced by Perplexity, which searches the web. A score for on-page AI optimization showed no link at all.

Do you need a Wikipedia page first?

No, and on its own it seems to add little once a brand’s wider prominence is counted. Kumar suggests small brands invest in Wikipedia, mainstream press and sustained YouTube presence, though he frames this as a hypothesis. Our data tests the Wikipedia part.

Brands with a Wikipedia article were more often named in the pages assistants cited, 74.3% against 59.5%. But once named there, they were recommended at the same rate, 50.2% against 48.9%. Most of the Wikipedia advantage reflected how well known those brands already were.

Consensus without an article also happens. Of the 110 options all four assistants named for a question, 21 had no Wikipedia article for themselves or a parent brand. In Kumar’s data, Wikipedia was also cited less often than video, media or forums, at 2.6% of citations. Coverage, not the encyclopedia entry, appears to do the work. Our guide on why Wikipedia matters for AI search covers where it does carry weight.

Can clear facts help a small brand beat a bigger one?

Yes, when the assistant can see them side by side. Chu and Hou (opens in a new tab) gave three AI models lists of skincare products with one real brand and nine invented ones. When everything was identical, the real brand always won. Our guide on whether AI assistants favor big brands gathers the wider evidence.

The moment an invented brand had a visible edge, the picture flipped. It won half the time with a rating 0.075 stars higher, 1.6 times as many reviews, or a price 7.3% lower. In a small follow-up where the products were first found by a simple search step, the real brand ranked near the bottom on relevance and was recommended 0% of the time.

The lesson for a small brand is to make real advantages explicit and comparable: ratings, review counts, prices, specifications and certifications. The caveat is that these tests handed the facts to the assistant directly. In real answers, your facts must first be on pages it finds.

Should a small brand win its home market and niche first?

The evidence points that way. A study of 66 European brands across twelve languages by Żatuchin (opens in a new tab), who is affiliated with the AI brand-intelligence company Rankfor.AI, found query language changed which brands were recommended far more than how they were described. Asking in a brand’s home language raised how often local champions were named by 0.80 on a 0 to 1 scale, against 0.15 for global multinationals.

In other words, a local champion that seems invisible in English answers may be the default recommendation at home. Check AI answers in the languages and markets your buyers actually use before concluding you are absent.

Specific buyer needs work the same way. Malthouse and colleagues (opens in a new tab) found that detailed questions about a buyer’s goals brought previously omitted brands into AI answers, and advise brands to stand for a few clear points of difference consistently everywhere they are described. The same team also looked at advertising, as our guide on whether ad spend helps AI recommendations explains.

Do shortcuts like bold claims or self-ranked lists work?

Briefly at best, and they carry real risks. Chu and Hou found that invented “clinical” claims let an unknown brand break through. But when every competitor used the same language, the advantage nearly vanished and the famous brand won again in 93.8% of trials.

Brands that did not compete that way fared worst of all: across 4,745 trials, they received zero recommendations. The authors used fabricated claims only to find the upper limit, and class them as potential false advertising. Real, verifiable evidence is the defensible version.

Self-ranked lists are common too. In our study of self-promoting lists, 24.2% of the numbered “best” lists that AI engines cited ranked their own publisher first. We found no detectable difference in how often their top pick was named compared with independent lists, so they are not a proven shortcut.

What should you do about it?

Build presence where AI assistants look, and make your advantage easy to check.

  1. Earn independent coverage. Pursue reviews, comparisons and inclusion in credible “best of” lists in your category.
  2. Show up where buyers talk. Genuine community discussion went with visibility in one study, and video was the top non-corporate source in another; fake posts do not count.
  3. Publish verifiable facts. Put prices, ratings, review counts, specifications and real certifications in plain, comparable form.
  4. Own a niche and a market. Measure AI answers for the specific needs, languages and places you serve before chasing the whole category.
  5. Keep the entity record tidy. Consistent names, websites and descriptions prevent confusion, even if they do not win recommendations alone.
  6. Measure repeatedly. Track unbranded questions across several runs and assistants, and expect progress in months, not days.

If you want help planning this, see our approach to generative engine optimization.

What does the research not tell us yet?

The research describes where small brands stand far better than it proves how to move them.

  • No controlled test shows a lasting lift. The tier advice is the vendor’s own hypothesis, and its planned experiment has not reported results.
  • Coverage findings are associations. Our data cannot show that new coverage causes new recommendations.
  • Lab tests simplify reality. The skincare experiments put the products in the question rather than letting assistants search.
  • Several sources have commercial ties. The tier and language studies come from authors tied to companies that sell AI visibility tools.

Frequently asked questions

Can a small business get recommended by ChatGPT?

Yes, but it starts behind. In one vendor’s data, small brands appeared in 9.9% of unbranded ChatGPT answers, so independent coverage and clear facts matter more than for big brands.

What is the fastest way for a small brand to appear in AI answers?

There is no proven fast route. The strongest signal in our data was independent coverage, and the fastest-rising small brands in one vendor’s data gained only 10 to 20 points across a three-month window.

Does a small brand need a Wikipedia page for AI visibility?

No. In our data, 21 of the 110 options that all four assistants named had no Wikipedia article for themselves or a parent brand.

Do fake reviews or bold claims help with AI recommendations?

Made-up claims swayed AI models in lab tests, but the edge vanished once rivals copied them, and invented evidence is potential false advertising.

Sources

Free strategy call

Some questions are easier to answer about your own business.

Bring the one that matters most. On a free 30-minute call we’ll take a first look at it and send you a short written read afterward.

Ask about my businessHow we work