Press and data

If you are putting together a piece on AI search visibility, this page exists to save you an email. Everything here is copy-paste ready, every number links to its method, and the last section is an offer that costs you nothing.

The one-liner, if you just need a sentence

Pick whichever length fits.

• 7 words: ShowUp Labs measures whether AI recommends you.
• One sentence: ShowUp Labs is an AI search visibility company that measures whether ChatGPT, Claude, Perplexity and Gemini name a business when its buyers ask, then does the technical and off-site work that changes the answer.
• With the differentiator: ShowUp Labs measures whether AI engines recommend a business, and publishes its own score daily, including when it is bad.

Numbers you can cite, and how each was produced

All of these come from measurement we ran and published, not from an estimate. The full per-site table is in the study so any row can be checked.

• 42 well-known sites averaged 80/100 on technical SEO and 51/100 on AI visibility. A 29-point gap. Only 6 of the 42 scored 70 or above on AI visibility. [Figure withdrawn 26 Aug 2026 pending re-measurement, why.]
• Canva scored 43/100 on technical SEO, near the bottom of the sample, and was named in 6 of 6 buying questions. Page set to noindex, content only rendering after JavaScript, no sitemap, no schema, no meta description.
• GoCardless scored 90/100 on the same technical checks and was named in 0 of 6. Stripe took 100% of those same payment answers.
• Roughly 85% of the sources AI engines cite are third-party pages rather than the brand's own site.
• 43 distinct third-party sources decide the answers to the 6 main buying questions in our own category, and we are named on none of them. We measure that weekly and publish the number whichever way it goes.

Method, stated plainly: technical checks are direct fetches and HTML parsing, not model opinion. The visibility half generates the real buyer questions for a category, puts them to a live web-grounded answer engine, and records which businesses get named. A check we cannot measure is left out of the score rather than counted as a failure.

Limitations, because you should print them: n=37 is a small sample, measured on a single day, against one web-grounded answer engine rather than polling ChatGPT, Perplexity and Gemini separately. Answers vary between runs and between engines. And we sell a product in this space, which is exactly why the full table is published rather than summarised.

Our own score, which is the part most vendors will not give you

Everyone selling AI visibility can measure themselves in about a minute. Very few publish the result. Ours updates daily at /proof and it is currently readiness 100, outcome 29, named in 0 of 6.

The outcome half is poor and the reason is the same one we tell clients: on-site work sets your ceiling and off-site presence decides where you sit under it. Our on-site work is finished and our off-site presence barely exists yet, so the number is exactly what the mechanism predicts. If that is a useful thing for your piece to point at, point at it.

A question worth asking every vendor you cover: what is your own AI visibility score, and what is the date you measured it? It sorts this market faster than any other question, and there are no qualifications in this category to sort it for you.

Free data for your article, no conditions

This is the useful part. Tell us the domains your piece covers and we will scan them and send you the results. Per domain you get the technical score, the AI visibility score, the buyer questions we asked, and which companies the engine named instead.

• It is free, and there is no agreement to sign.
• You are not obliged to mention us. If the data is useful and the piece is better without our name in it, run it without our name in it.
• We will tell you when a result is unreliable rather than send you a confident number, which happens when a site blocks our fetcher or the grounding search gets rate limited.
• If you want to check us first, run the free scan on your own site. Takes about a minute.

Ask via the contact form. The reason we do this is not complicated: 85% of what AI engines cite is somebody else's page, we are on almost none of them, and being genuinely useful to the people writing those pages is the honest way to change that.

If we have written about your company

We publish automated, dated measurements of public websites at /ai-visibility/, the same shape as a PageSpeed or Lighthouse score. Every page states the date, states that it was automated, and describes the method. We measure whether engines can read and cite a site. We do not rate the business.

If you work at one of them and think something is wrong, tell us and we will re-run it. We would rather be corrected than be confidently wrong, and that offer is on every published report.

Details

• Name: ShowUp Labs (plural). Not ShowUp Lab, an unrelated event operations platform in Bengaluru, and not ShowUpWithAI. All three appear in the same searches.
• Founded by: Aaron James. Part of Cutworks Studio.
• What it costs: the scan is free with no signup. The full report is £39.
• Logo and brand: the mark is at /logo.png. Ask if you need it another way.
• More on the method: about and the full method.

Our crawler

If you have seen ShowUpLabsBot in your logs, showuplabs.com/bot says exactly what it fetches, how often, and the two robots.txt lines that stop it. It obeys robots.txt, checks before fetching rather than after, and does not train models on anyone's content.

Questions, answered