When someone asks AI for a clothing brand does it say your name?
Find out in 3 minutes. Then get the list that puts your name in the answer.
"Best X for Y" is the shape of almost every shopping question. One answer, three names.
Free: one check per email. No card, about 3 minutes, and a password to see your report. You read the real AI answers, word for word.
One real check: TuAbogadoColombia, a law firm in Bogota
From the check of 30 Sept 2026. ChatGPT's answer, word for word; the question was asked in Spanish.
The question
“¿Cuál es el mejor abogado en Bogota?”(Who is the best lawyer in Bogota?)
- 1.Brigard Urrutia
- 2.Posse Herrera Ruiz
- 3.Gómez-Pinzón
- 4.Philippi Prietocarrizosa Ferrero DU & Uría (PPU)
- 5.Baker McKenzie Colombia
TuAbogadoColombia is not in the answer. The rivals are.
- 1.Brigard Urrutia
- 2.Posse Herrera Ruiz
- 3.Gómez-Pinzón
- 4.Philippi Prietocarrizosa Ferrero DU & Uría (PPU)
- 5.TuAbogadoColombia
Not a measured answer. It shows where the work aims.
Today
Estimate: every open task done
How Rankotan works
Four steps, and a report at the end of them you can check line by line. Then, for the people who want it, the method underneath.
1Rankotan reads your site
The same way an AI crawler does — the HTML before the JavaScript, your robots.txt, your structured data, the pages you actually sell from. That crawl is 32 of your 100 points, and every finding comes back as a sentence you can verify in your own browser.
2Rankotan asks what your customers ask
Real buying questions for your trade and your area — not a keyword list. The ones at the top get asked three times, because AI answers move between runs and one sample is not a measurement.
3Rankotan shows you the answers, word for word
Your position in every answer, who was named instead of you, which KIND of question you lose, and every page the engines cited to decide. Nothing summarised away, nothing you have to take on faith.
4Rankotan's team does the work
A prioritised list where every hour is booked against a source in your own report. Do it yourself from the same list, or hand it to us and watch the number move when you scan again.
We ask ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews, live, with search grounding
Every answer carries the sources behind it — no cached model knowledge, no guessing what AI said last year. Five engines today, named plainly rather than implied — and they disagree violently, which is the whole point: on one business, on one day, Perplexity named them in 79% of answers, Gemini in 46%, and Google's AI Overview in 7%. Your report shows a column for each, so you can see exactly what is covered and what is not.
- Perplexity
- Gemini
- Google AI Overviews
- ChatGPT
- Claude
Sources cited
- tripadvisor.commentions you
- reddit.comnever mentions you
- timeout.comnever mentions you
- booking.commentions you
- 369
- businesses ranked by AI across our reference set
- 280
- sources the engines cited — all recorded, all checkable
- 120
- real buying questions asked to build the sample reports on this site
Our own numbers, from the reports published on this site. We would rather show you a small true number than a large invented one.
Being named and being quoted are not the same thing. AI can describe your business without quoting your site, and quote your site without naming you. On Gemini, as few as 30% of the brands mentioned are also cited as sources — so we measure both.
Semrush & Adobe, AI Visibility Index 2026
The method
- AI SEO, AEO and GEO
- Three names for the same thing. AEO is answer engine optimisation, GEO is generative engine optimisation, and AI SEO is what everyone actually calls it. It is not ranking in a list of links — it is being named inside the answer.
- 15 questions, free
- Real buying questions for your trade and your area, generated per language — not a generic keyword list. The main ones run three times on each conversational engine, because AI answers vary between runs and one sample is not a measurement. Fifteen questions is roughly sixty-five live API calls. The paid tiers ask the full set and add branded questions on top.
- Composite score out of 100
- Findability 45 · Readiness 32 · Answerability 15 · Presence 10. Those weights are fixed and published, so two scans are comparable. Together the four parts make 102 measured points; the score you see is out of 100: points earned ÷ points measured × 100, so a part we could not measure is left out of the division, never counted as zero.
- Re-scan after the fix
- AI answers drift. One snapshot tells you almost nothing; a re-scan after you fix something tells you whether the work is working.
- Every raw answer stored
- We keep the full response behind every score. When you ask why a number moved, there is a real sentence to point at.
Rankotan does the work, not just the report
Most tools hand you a dashboard and wish you luck. Behind this one is a team that actually does the fixing: the citations, the Google profile, the reviews, the pages that get quoted.
We work in English and Spanish because your customers do. And we tell you plainly when something is not worth doing.
— The Rankotan team
One scan is free. Fixing it is $79 a month — you do the work, we make it easy.
Free: one scan per email shows who AI names instead of you, what it cited, and the three things worth fixing first. Paid, $79/mo: re-scans whenever you ask, to check your fixes, and every gap turned into a job written for your business — why it matters, the steps, and text ready to paste. Want it done for you instead? Expert hours are a paid add-on, below.
The scan and the report
Everything we measured, yours
Free
- The verdict, and every question we asked
- Who was named instead of you, and on what kind of question
- Every raw answer, and the full citation ledger of pages AI cited
- How many fixes are open, and what they are worth
- No card, no trial that expires — just a password for your account
Monitor + Plan
We watch and write the plan; you fix
$79/mo
- Everything in the free scan
- Re-scans whenever you ask, on the same questions — your account shows how many you have left
- A written job for every open gap — steps, text ready to paste, and a way to check it
- Competitor tracking
- The monthly plan never does the work for you — expert hours are the add-on for that
The paid add-on
Want it done for you instead? Hire the hours.
The monthly plan never does the work — it turns every gap into a job you can do yourself. If you would rather hand that job to someone, expert hours are a paid add-on: 20, 40 or 60 hours one-on-one with a hired SEO expert. No price is fixed here — every request starts as a quote, answered by Ignacio himself.
Expert hours: 20
20 hours one-on-one with a hired SEO expert, working through your open jobs — citations, your Google profile, schema, reviews, content and your phone listing.
Expert hours: 40
40 hours one-on-one with a hired SEO expert, working through your open jobs — citations, your Google profile, schema, reviews, content and your phone listing.
Expert hours: 60
60 hours one-on-one with a hired SEO expert, working through your open jobs — citations, your Google profile, schema, reviews, content and your phone listing.
What we will always do
Four promises. Each one is checkable, which is the point.
You see the actual answers
Word for word, engine labelled. Not a score you have to take on faith — the sentences AI wrote about your trade, in full.
Our formula is published
The whole scoring method is public: what we weight, why, and how we count. Check our arithmetic. Most tools will not show you theirs.
English and Spanish, scored apart
Two languages, two question sets, two scores. Being visible in English and invisible in Spanish is not a rounding error in Miami — it is half your market.
Every fix traces to a real gap
We only recommend something when we can point at the source that cites a competitor and not you. No keyword stuffing, no busywork billed by the hour.
The market figures on this page come from
- Semrush & Adobe
- SOCi
- BrightLocal
And from our own scans, published on this site
- 369
- businesses ranked by AI
- 280
- sources the engines cited
- 120
- real buying questions asked
No testimonials, because we have no customers yet. We would rather show you a small true number than a large invented one.
Before you hand over an email
The questions we actually get asked
- Is this just SEO with a new name?
- No, and the difference is measurable. SEO asks where you rank on a page of ten blue links. This asks whether an assistant NAMES you when somebody describes their problem out loud — and there is no second place in an answer. A site can rank well and be absent from every answer about it; we have measured exactly that.
- Where does the number come from?
- Fifteen questions your customers would actually ask, put to every assistant we can reach, with the top ones asked three times because AI answers move between identical runs. Every raw answer is stored, so the score can be rebuilt from the evidence. The report prints the questions and what came back — you never have to take the number on faith.
- Why is my score low when my website is fine?
- Usually because being found and being readable are different problems. A site can be beautiful, fast and completely invisible to something that has to quote it — and most of what decides whether you get named happens off your own site, on the pages the assistant actually cites. The report separates the two so you know which one you have.
- Do you make things up if the AI does not know my business?
- No. We refuse to store an answer that came back without search grounding, because an ungrounded model invents local businesses that do not exist — we have watched it name plumbers that were never founded. Discarded answers are left out rather than counted against you, and the report says how many there were.
- What happens to my phone number?
- It is used to talk to you about your own results, including a call from Rankotan, which says it is an AI at the start. We never sell your details, pass them to advertisers, or add you to unrelated lists. One reply saying stop is enough, by any route.
- How long does the scan take, and what does it cost?
- A few minutes — usually under fifteen — and nothing. One free scan per email, ever. You confirm your email first, because sixty-odd questions to five engines is real money and we would rather spend it on people who want the answer.
- Will the score go up if I do the work?
- That is the part we are building next and the part we will not overclaim today. What exists now is the measurement and the list of what is missing, with the evidence beside each item. Re-scanning shows the change; nobody can promise a particular number.
RankotanFind out what AI says about you
One field, no card. Set a password and you'll see the actual answers.
Free: one check per email. No card, about 3 minutes, and a password to see your report. You read the real AI answers, word for word.
