How to find the competitors ChatGPT recommends instead of you

Ask ChatGPT the questions your buyers ask, several times each, and count every brand the answers name, once per answer. The brands named most are the competitors ChatGPT recommends, whether or not you'd call them rivals. Your share of voice is your mentions over yours and theirs together.

Checked against the studies and help pages it cites on October 20, 2026.

The method

Find them in five steps.

Checking whether ChatGPT names you is one count. Finding who it names instead is a tally of every brand in every answer. Set up the chats the way the guide to checking ChatGPT by hand describes, then count like this.

  • 1. Ask for products, not about yourself

    Write the questions a buyer types before they know you: your category (“best privacy-friendly analytics”), a need (“analytics that won't slow my site”) and alternatives to the market leader. Each kind of question can surface a different set of rivals.

  • 2. Ask each one many times

    In SparkToro's study of 2,961 runs, the chance that ChatGPT gave the same list of brands twice was under 1 in 100. The study ran each prompt 60 to 100 times, and its author suggests at least that many to know an AI's set of recommendations.[1] An audit of product advice found the same: the products recommended often change across repeated requests.[2] By hand, a few runs per question at least separates the brands that keep coming back from the ones that turned up once.

  • 3. Count each brand once per answer

    List every product the answer offers as an option, and give each a mark per answer, however often the answer repeats it. Merge spellings: in our count, ChatGPT wrote “Fathom Analytics” in 37 answers and “Fathom” in 7, one product either way. Leave out your own names, and keep a column for answers that name you.

  • 4. Note who comes first

    Mark the brand each answer offers first. A rival named in most answers but rarely first is a different threat from one that leads when it appears.

  • 5. Work out your share of voice

    Pick the rivals that recur, then divide your mentions by your mentions plus theirs. It reads differently from your mention rate: in our example Plausible was named in 80% of answers but held 20% of the voice, since most answers named several brands.

Our data

Who ChatGPT named for an analytics tool's buyers.

What a tally looks like: the brands ChatGPT named when we asked Plausible's buyer questions, each counted once per answer.

32

other brands were named across 65 answers, 17 of them in only one answer.

42 of 65

answers named Matomo, the brand named most after Plausible: 55% to 74% at 90%.

20%

share of voice for Plausible among its tracked competitors (17% to 25%), though answers named it 80% of the time.

Share of 65 ChatGPT answers naming each brand
Plausible80%
Matomo65%
Umami63%
Fathom Analytics57%
Simple Analytics43%
PostHog32%
Google Analytics31%
GoatCounter15%

25 more brands were named in fewer answers.

65 ChatGPT answers to the 22 buyer questions GEOFast tracked for its plausible.io demo project, in 3 runs within 62 minutes on October 6, 2026, read from the ChatGPT app signed out in US English. Questions about the brand itself are left out. Brands are the products each answer offered as options, read by a model and counted once per answer; spellings that differ only in case, punctuation or a trailing word such as “AI” or “Inc” count as one. Ranges are 90%.

Limits: one field and runs within about an hour, so this shows who leads in one market on one day, not a trend. Share of voice is among the 10 competitors the project tracks.

In GEOFast

How the X-ray and Competitors count it.

GEOFast does the asking and the tally for you, from a first look to a tracked measure.

  • “Named instead of you”

    When the free X-ray's site checks end, it asks ChatGPT, Perplexity and Gemini three of your buyers' questions once each and tallies the brands their answers name, each once per answer, your own names left out. Without an account you see the four named most and a count of the rest; the heading reads “Named alongside you” once an answer names you. Try it with the AI visibility checker.

  • A free account's first check

    Six of your questions, once each on the same three assistants: 18 answers, with every brand they name and the sources they cite. Still a first look: too few answers to hold a rate steady.

  • Competitors, tracked

    The Competitors page ranks every tracked brand by how often it's named, how often it's the first pick and its share of voice, with a 90% range, run by run. “AI also names” lists the brands the answers name that you don't track, and “Where they beat you” lists the prompts where a competitor is named and you aren't. On Growth, Agency and the trial.

  • Why they win

    The Sources page lists the pages the answers cite and marks the ones that name a competitor and not you: the pages to get onto.

FAQ

Questions about AI competitors.

  • Can I just ask ChatGPT who my competitors are?

    You can, but it answers a different question: what it knows about your company. The competitors that cost you sales are the brands it names when a buyer asks for a product like yours without naming you, so count those answers instead.

  • Are the brands ChatGPT recommends my real competitors?

    They are the brands your buyers hear about beside you, which can include products you'd never list as rivals: bigger suites, free tools, other kinds of product. Track the ones that come up across several questions.

  • Why does ChatGPT recommend a competitor for one question and not another?

    Each question leads to different searches and different pages, and each answer is written afresh. A brand that leads across most questions competes with you everywhere; one that leads on a single question competes on that need.

  • What's the difference between mention rate and share of voice?

    Mention rate is the share of answers that name you. Share of voice is your mentions against yours and your tracked competitors' together. In our example, ChatGPT named Plausible in 52 of 65 answers (80%), yet its share of voice among 10 tracked competitors was 20%, because most answers named several brands.

Sources

Where each fact comes from.

Each numbered source is the page a fact comes from: published studies and the AI companies' own help pages, read on October 20, 2026. The answers above repeat facts marked in the guide. Help pages change, so check them before you act.

  1. 1SparkToro: NEW Research: AIs are highly inconsistent when recommending brands or products; marketers should take care when tracking AI visibilitysparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-brands-or-products-marketers-should-take-care-when-tracking-ai-visibility/
  2. 2Uberti-Bona Marin et al., arXiv: “If I Had to Buy Just ONE: Galaxy S26 Ultra”: Auditing AI-Generated Product Recommendationsarxiv.org/abs/2609.18729

Keep reading

Free X-ray

Find out what AI says about your site.

Paste your URL and leave with a ranked list of what to fix, with steps and code.

  • Free
  • No account