What AI visibility is, and why a Spanish brand cares now
AI visibility is how often, and in what terms, your brand appears when someone asks ChatGPT, Gemini, Perplexity or Google AI a question about your category. "Which Barcelona clinics are good for sensitive skin." "A boutique hotel in Girona for a long weekend." "An estate agent that works with foreign buyers in Valencia." The answer names three or four brands. If yours is not one of them, you did not lose the pitch. You were not in the room.
Three numbers explain why this matters for a brand in Spain this year, not in some future one:
- Regular ChatGPT use in Spain went from 4% in 2023 to 28% in 2025 (Funcas, III Encuesta sobre IA, January 2025). Four in ten people under 45 use AI tools daily.
- Organic search's share of traffic to Spanish websites fell from 52% to 41% in a single year (SE Ranking, Tráfico de IA en España, June 2026, 101,574 sites).
- 84% of the sources AI assistants cite are earned media: journalism, industry press, reviews. Paid content gets 0.3% (Muck Rack, What Is AI Reading?, May 2026, 25 million links).
The third number is the useful one. The sources assistants trust most are exactly the ones a decent PR effort already produces. What most brands lack is measurement and structure: nobody knows the current share of voice, and the website does not tell machines who the brand is, in a format they can read. We do both, and we report it monthly.
Two products: the audit first, then the program
| AI Visibility Audit | AI Visibility Program | |
|---|---|---|
| Time | 5 days | From 3 months |
| Scope | 30 category questions × 4 engines × 5 competitors | A monthly cycle: content, structure, PR alignment, measurement |
| Engines | ChatGPT, Gemini, Perplexity, Google AI | The same four |
| You get | Share of voice against five competitors, the sources each engine cited, a technical check of your site, a ten-step plan | Citable content every month (guides, data, comparisons), entity structure on your site, your PR calendar aligned with the questions people ask, a share-of-voice report with confidence intervals |
| Price | Fixed fee, quoted within 24 hours | Monthly fee, quoted within 24 hours |
| Best for | Knowing where you stand before spending anything else | Moving the number, month after month, with someone accountable for it |
The audit stands on its own. Some brands run it once, fix the ten things on the list, and come back a quarter later to measure again. That is a perfectly good use of it. The program is for brands that want the number to move and a report they can show the board.
What a share-of-voice report shows
Every report has the same spine: for each engine, how often each brand was named across the 30 questions, with a confidence interval, and which sources the engine leaned on. Here is the shape of it, with worked example numbers.
| Brand (example) | Mention rate, ChatGPT | 95% confidence interval | Sources cited most |
|---|---|---|---|
| Your brand | 23% | 17% to 30% | La Vanguardia, your ingredients guide, a dermatologists' ranking |
| Competitor A | 41% | 34% to 48% | Wikipedia, El País, Trustpilot |
| Competitor B | 30% | 24% to 37% | Industry press, own blog |
| Competitor C | 12% | 8% to 18% | Reddit, a directory |
| Competitor D | 9% | 5% to 14% | Own site only |
| Competitor E | 4% | 2% to 8% | Not cited; named from training data |
Three things to read in your report:
- Mention rate per engine, not a blended score. Engines barely overlap in what they cite, so a single "AI visibility score" hides more than it shows. You see four columns, one per engine, plus the position your brand tends to take when it is named.
- Citations and mentions, separately. About 62% of AI citations are "ghost citations": your URL is used, your name is not said (Semrush, June 2026). Both are tracked, because they need different fixes.
- The source list is the to-do list. The third-party pages that drive your competitors' mentions are where your PR and your content should go next. Every audit ends with that list.
How it works, week by week
Week 1: the audit. Day 0 is a 45-minute brief: your category, your five competitors, what you already publish. Days 1 and 2, we build the 30 questions from the buyer journey (definition, comparison, alternatives, use case, purchase) and run them in clean sessions across the four engines, repeated so a single odd answer cannot skew the rate. Days 3 and 4, scoring, source analysis and the technical check: can OAI-SearchBot and PerplexityBot read you, do your pages carry entity and schema data, is there anything on the site worth citing. Day 5, the report and the ten-step plan, on a call.
Week 2: decisions. If you continue into the program, we agree the plan, the monthly content slots and how your PR calendar feeds in. The fast fixes ship this week: entity structure, schema, the pages that say plainly who you are.
Weeks 3 and 4: first citable pages. Guides, data, comparisons, whatever the audit showed the engines are hungry for in your category, in Spanish and English. Your PR agency gets the list of questions to aim the next releases at.
Every month from then on: publish, align, measure, report. The report lands with the same 30 questions, so month three compares cleanly with month one. At month three we review the plan together and decide what to do with the next quarter.
What we commit to
- Rates with confidence intervals. Every number comes from repeated runs, with the interval printed next to it, and we say plainly when a month's change sits inside it. A one-run dashboard is a horoscope. A rate is a measurement.
- Presence on the shortlist, and the number that proves it. AI referrals are 0.30% of visits to Spanish sites today (SE Ranking, June 2026), so what you are buying is being named when a buyer asks "which one": better-qualified leads and a metric you can track. If someone promises sessions, ask them where from.
- A realistic clock, in an open market. First citations on long-tail questions are plausible in 4 to 8 weeks; contested category questions take months and press coverage naming you. Only 15% of ChatGPT categories have a clear owner, and owners keep their spot 90% of months (Semrush, July 2026). First in with depth tends to stay.
- Work on what is actually controllable. Nobody controls a placement in ChatGPT: Google AI Mode replaces 56% of its cited sources every week, and after three runs of the same prompt only 2.2% of ChatGPT citations persist (Kevin Indig, Search Engine Land, June 2026). Anyone selling "guaranteed in ChatGPT" is selling a screenshot. We work on what does move the rate: what you publish, how your site is structured, and who cites you.
Who it is for
Brands in Spain and Europe whose customers ask an assistant "which one" before they ask Google: beauty and aesthetic clinics, hotels and restaurants, real estate, wellness, and B2B services where the buyer shortlists three suppliers in a chat window. The program fits best where there is already some press activity to align, or a plan to start it.
PR and communications agencies: this is a product you can sell tomorrow to the brands you already run coverage for. Your publications plus our measurement and structure is the whole thing. The referral, co-branded and white-label models, and a free audit for one of your clients as a first-meeting gesture, are on the partners page.
Method, glossary, data, sources
We publish how we measure, because a number without a method is an opinion.
- How the audit works, question by question: the 30-question design, the four engines, repeated runs, the share-of-voice formula, a sample report.
- Glossary: GEO, AEO, LLM SEO, share of voice, ghost citation, query fan-out, entity, AI Overviews vs AI Mode, in English and Spanish.
- AI visibility in Spain: the numbers: every Spain figure we use, with source and date, refreshed as new data comes out.
- Not sure the answer box is your first problem? The Research Sprint reads your market in five days and takes a light measurement of how the assistants mention you and five competitors on the way.
- Sources: the studies behind every claim on this site.
Everything above also exists in Spanish: visibilidad en IA, metodología de la auditoría, glosario, datos de España and fuentes.
Ready when you are. Tell me your category and five competitors, and you will have a fixed fee for the audit within 24 hours. If you want a page to catch what the visibility brings, the Launch Page takes 72 hours.