Measured · 2026-08-07

Who ChatGPT and Gemini recommend for test automation software

We put 6 buyer questions about test automation software to ChatGPT and Gemini. Eight tools tied at the top, each named in 2 of 6 answers, and none in all six. Here is the full list, the pages the answers were built from, and what it means whether or not your product is on it.

Answers read
6
Products named
33
Top source
bugbug.io

Eight tools tied for the most mentions in our test automation software measurement, each named in just 2 of the 6 buyer questions we ran on ChatGPT and Gemini on 2026-08-07: ACCELQ, BrowserStack, BugBug, Cypress, Katalon True Platform, mabl, Playwright and Selenium. No product did better than 2 of 6.

Across the six questions, ChatGPT and Gemini named 33 different products a combined 41 times. 25 of those 33 were named only once, or in 1 of 6 answers (17%). The headline for a marketer is that this category has no consensus pick: a single mention is by far the most common outcome, and the top of the list tops out at a 33% hit rate.

We ran six phrasings of the same buying question, each as one live web search per assistant: "best test automation software," "best test automation software for small teams," "what test automation software should I use," "best free test automation software," "most affordable test automation software," and "test automation software recommendations." Because the leading eight tools all sit at 2 of 6, there is no default answer here, and the shortlist a buyer sees swings hard on which of the six phrasings they type.

How the shortlist changed from "best" to "free" to "affordable"

The bare "best test automation software" question returned an enterprise-leaning set of 5: mabl, Katalon True Platform, Perfecto by Perforce, Keysight Eggplant and TestRail. Three of those, Perfecto by Perforce, Keysight Eggplant and TestRail, appeared in this question and in none of the other five.

"best free test automation software" returned a completely different world: 8 open-source frameworks, namely Playwright, Cypress, Selenium, Puppeteer, WebdriverIO, TestCafe, Taiko and Gauge. Five of them, Puppeteer, WebdriverIO, TestCafe, Taiko and Gauge, appeared only in the free question. The single word "free" flipped the answer from commercial platforms to developer libraries.

"most affordable test automation software" returned yet another set of 6, and it barely resembled the free list: EZ Testing, BugBug, Xping, TestPlan, Gondola and TestFiesta. Five of these, EZ Testing, Xping, TestPlan, Gondola and TestFiesta, appeared only in the affordable question.

The most important number in the whole measurement: the "best free" set and the "most affordable" set share zero products. Free returned open-source tools, affordable returned small paid vendors, and not one tool landed on both lists. If you treat "free" and "affordable" as the same buyer, the assistants disagree, they answer them as two separate questions.

"what test automation software should I use" acted as the hub, with 8 names: Playwright, Selenium, Cypress, mabl, Testim, ACCELQ, Tricentis Tosca and BrowserStack. It borrowed mabl from the "best" set, Playwright, Selenium and Cypress from the "free" set, and ACCELQ and BrowserStack from the "recommendations" set. Only Testim and Tricentis Tosca appeared here and nowhere else.

"test automation software recommendations" returned an API- and platform-leaning set of 8: BrowserStack, Postman, Salesforce Platform, ACCELQ, Apidog, QA Wolf, Qase and Testlio. Six of them, Postman, Salesforce Platform, Apidog, QA Wolf, Qase and Testlio, appeared only in this question. Only ACCELQ and BrowserStack carried over, both shared with the "what should I use" answer.

"best test automation software for small teams" returned 6: Endtest, BugBug, TestAgent, Katalon True Platform, EirTests and TestWheel. Four of them, Endtest, TestAgent, EirTests and TestWheel, appeared only here, while Katalon carried over from "best" and BugBug reappeared in "affordable."

Where the answers came from

The 41 product mentions rested on 45 citations across 37 domains. The two most-cited domains were bugbug.io and pie.inc at 3 citations each, followed by contextqa.com, github.com, katalon.com and tinyctl.dev at 2 each. Every other domain was cited once.

Independent roundup blogs did a lot of the work. pie.inc fed three separate answers (small teams, what should I use, and recommendations), while contextqa.com and tinyctl.dev each fed two. Because "what should I use" and "recommendations" drew on the same three roundups (contextqa.com, pie.inc and tinyctl.dev), the same two products, ACCELQ and BrowserStack, are the ones bridging those answers. Shared sources produce shared names.

Here is the key insight for a vendor: in this category the assistants cite vendors' own websites, and those vendors get named. 18 of the 33 named products had their own domain appear as a source, including BugBug via bugbug.io (3 citations), Katalon True Platform via katalon.com (2), ACCELQ via accelq.com, mabl via mabl.com, TestRail via testrail.com, Perfecto by Perforce via perfecto.io and Keysight Eggplant via keysight.com, down to small names like Xping (xping.io), Gondola (gondolatest.com), TestPlan (test-plan.io), TestFiesta (testfiesta.com) and EZ Testing (ez-testing.com). Your own site is not too promotional to be cited here, it was the most common source type in the whole set.

Open-source tools got in through GitHub: github.com was cited twice, feeding the "best" and "best free" answers, which is how Playwright, Selenium, Cypress, Puppeteer, WebdriverIO and TestCafe reach the shortlist. If your project lives on GitHub, that README is a page the assistants are reading.

The big review directories were nearly absent. G2 was cited once (learn.g2.com) and Gartner once (gartner.com, in the bare "best" question only). Capterra did not appear at all, and neither did any Reddit or Quora thread. Stack Overflow showed up only as its company blog (stackoverflow.blog), once. In a category people assume is ruled by review grids, G2 was cited exactly 1 time out of 45.

What a vendor here would do to get named

Getting named is a sourcing problem, not a market-share problem. 5 of the 6 products in the "most affordable" answer, EZ Testing, Xping, TestPlan, Gondola and TestFiesta, are small vendors that got named because their own domain was the cited source. No brand recognition put them there, a page the assistant retrieved did.

So the play is concrete: be present in the exact sources these answers used. Publish the pricing and comparison pages on your own domain, the source behind 18 of the 33 named products. Earn a place on the independent roundups the assistants lean on, led by pie.inc (3 answers), contextqa.com and tinyctl.dev (2 each). Keep your open-source repo current on GitHub. Chasing a G2 badge barely moves these particular answers, where G2 was cited once in 45.

Pick your phrasing battles too. Because "free" and "affordable" returned zero shared products, a free tool that wants the "affordable" list, or a cheap paid tool that wants the "free" list, is fighting for a shortlist it is not on today. The 2-of-6 ceiling cuts the other way as well: no competitor has locked this category down, so the room at the top is real.

The full list, counted

We kept each product name exactly as the assistants returned it, so "Katalon True Platform" is listed as named rather than shortened. Below these eight, 25 more products were each named once, or 1 of 6 (17%).

ProductNamed inShare
ACCELQ2 of 633%
BrowserStack2 of 633%
BugBug2 of 633%
Cypress2 of 633%
Katalon True Platform2 of 633%
mabl2 of 633%
Playwright2 of 633%
Selenium2 of 633%

Where Glotier fits

Glotier does not sell test automation software, so we are correctly absent from all 6 of these answers. We would not fake our way onto a list we do not belong on, and the data shows we did not.

This measurement is, though, exactly what we run for the category you do sell. We put the six buyer questions to the assistants, record every product named and every source behind them, and show why the named products are named. On 2026-08-07 that was 33 products, 41 mentions and 37 source domains for test automation software, and for you it would be your product and your sources in those slots. The check is free, no account, and takes about a minute.

The questions we asked

One live web search per question, put to exa+or:chatgpt,gemini on 2026-08-07. 6 of 6 came back with an answer we could read. Whether a product was named is decided by looking for it in the answer text, not by asking a model for its opinion.

  1. best test automation software
  2. best test automation software for small teams
  3. what test automation software should I use
  4. best free test automation software
  5. most affordable test automation software
  6. test automation software recommendations

Questions people ask

What is the best test automation software according to ChatGPT and Gemini in 2026?
There was no single winner. Eight tools tied at the top, each named in 2 of the 6 buyer questions (33%): ACCELQ, BrowserStack, BugBug, Cypress, Katalon True Platform, mabl, Playwright and Selenium. The bare 'best test automation software' question specifically returned mabl, Katalon True Platform, Perfecto by Perforce, Keysight Eggplant and TestRail.
What is the best free test automation software?
The 'best free' question returned 8 open-source frameworks: Playwright, Cypress, Selenium, Puppeteer, WebdriverIO, TestCafe, Taiko and Gauge. Five of them (Puppeteer, WebdriverIO, TestCafe, Taiko and Gauge) appeared only in the free question, and this free list shared zero products with the 'most affordable' answer.
Is 'free' the same as 'affordable' when AI recommends test automation tools?
No. The free set and the affordable set shared 0 products out of the 6 questions. 'Free' returned open-source tools (Playwright, Cypress, Selenium and others), while 'most affordable' returned small paid vendors: EZ Testing, BugBug, Xping, TestPlan, Gondola and TestFiesta. ChatGPT and Gemini answer them as two different questions with two different shortlists.
Do G2 and Capterra decide which test automation tools AI recommends?
Not in this measurement. Out of 45 citations, G2 was cited once (learn.g2.com) and Gartner once, while Capterra, Reddit and Quora were not cited at all. The heavy lifting came from vendors' own domains, which fed 18 of the 33 named products, plus independent roundups like pie.inc, contextqa.com and tinyctl.dev.
How can my test automation tool get recommended by ChatGPT and Gemini?
Be in the sources these answers used. 18 of the 33 named products had their own domain cited, so your pricing and comparison pages matter directly. Independent roundups carried weight too (pie.inc fed 3 of the 6 answers), and open-source tools got in through GitHub. It is a sourcing game, not a market-share game: 5 of the 6 'most affordable' picks were small vendors named through their own website.

Do you sell in test automation software? Find out whether you are in that list.

Paste your domain and watch the same run happen for your own buyer questions: which of the three assistants names you, who gets named instead, and the exact pages those answers were built from. Free, no card, no account for the first check.

Show my visibility

For reference, ACCELQ was named in 2 of the 6 answers we read.

Start with Solo

Get Glotier Solo

Everything on this page is one measurement, taken by hand, on one day. Solo runs it for your product every day and writes the work it points to.

  • The Citation Agent: which of your pages an assistant can actually cite, and the fix for each
  • A paragraph-quality read of your pages: what an assistant can lift whole, and where a claim is missing its source
  • An X agent and a Reddit agent: the live threads worth answering in your category, with a reply drafted for each
  • A daily check, and an article written from what it measured that day, ready to publish
  • The same buyer questions re-asked every day, with every source page behind each answer

Track 3 products and up to 150 buyer questions across ChatGPT, Gemini and Perplexity. $39/month.

Cancel any time. Not for you? Email us within 7 days of a charge and we refund it in full. Refund policy

Not ready to pay? The check is free, with no card and no account. Run it on your own product.