El 91% de las PYMEs españolas no existe para ChatGPT
Four AI engines analyzed: ChatGPT, Gemini, Perplexity and Claude. 91% of Spanish SMBs only show up in one. But invisibility is not uniform: 1 in 4 hotels appears across multiple AIs, versus 1 in 50 auto repair shops. AI invisibility is a problem of businesses that customers don't need to physically visit.
Surfeo's AI Recommendation Gap study tested 9,865 businesses across 30 industries and 10 cities by querying ChatGPT, Gemini, Perplexity, and Claude directly. The finding: 91% of businesses are only recommended by a single AI model — leaving them invisible to customers using any of the other three.
What does 'found in just 1 of 4 AIs' mean? We asked each AI model to recommend businesses for a given industry and city. If a business was only mentioned by one model out of four, it means three-quarters of AI-powered search will never surface it. For a customer using any other AI, that business effectively doesn't exist.
Why this study matters now
Two years ago, a small business worried about showing up on Google. It was the only relevant entry point. That door is splitting and starting to open elsewhere:
ChatGPT use in Spain has multiplied by 7 in two years, from 4% to 28% of frequent use between 2023 and 2025 according to Funcas. The Spanish National Statistics Institute (INE) confirms that 37.9% of the population used generative AI in the last three months of 2025.
When Google shows an AI answer above the results, click-through rates to websites halve: from 15% to 8%, according to Pew Research analysis of 68,879 real searches. One in five Google searches now triggers that AI answer.
Gartner predicts traditional search volume will drop 25% by 2026 due to substitution toward AI chatbots. That decline is already documented in Pew and CNMC data.
76% of Spanish SMBs already use AI in their daily work according to the 2nd Hiscox SMB Report (December 2025). But almost none measure whether those same AIs recommend them to potential customers. AIs learn from what already exists digitally: businesses that position themselves first — directories, reviews, content, structured data — reinforce their position every time the model is retrained. Those who wait three years will find the space taken by competitors who acted.
How we ran the study
We queried ChatGPT, Gemini, Perplexity, and Claude about real businesses across 30 industries and 10 cities in Spain. 9865 AI responses were collected and analyzed to map the AI recommendation landscape.
“Five years ago you only needed to show up on Google. Today a customer can ask ChatGPT for a dentist in Seville and, according to our data, in 96% of cases no specific local dentist will appear. AI recommends what it already knows, and almost no Spanish SMB has done the work to make itself known to AI. The dangerous thing about AI invisibility is that it's silent: you only find out when customers stop coming.”
Industry breakdown
Taller mecánico have the widest AI gap: 98% are recommended by just 1 AI model
The industry data tells a clear story: businesses with extensive third-party digital footprints get recommended far more often. Hotels and restaurants benefit from aggregator platforms — Booking, TripAdvisor, Google Hotels, Yelp — that give AI models rich, structured data to draw from.
Professional service businesses face the opposite problem. A plumber, photographer, or mechanic typically has fewer digital touchpoints: perhaps a website and a Google Business listing. That's not enough for most AI models to confidently recommend them.
The five industries with the largest AI recommendation gap are all local services: auto repair, photography, hair salons, gardening, and carpentry. The pattern is consistent: fewer indexable data sources equals lower AI recommendation rates.
| Industry | Businesses | Avg score | Single AI only |
|---|---|---|---|
| Taller mecánico | 328 | 35/100 | 98% |
| Fotografía | 252 | 35/100 | 98% |
| Peluquería | 353 | 36/100 | 97% |
| Dentista | 394 | 36/100 | 96% |
| Jardinería | 331 | 36/100 | 96% |
| Nutricionista | 374 | 35/100 | 95% |
| Agencia de marketing | 404 | 36/100 | 95% |
| Gestoría | 366 | 36/100 | 95% |
| Fontanería | 333 | 36/100 | 95% |
| Carpintería | 368 | 36/100 | 95% |
| Clínica estética | 382 | 36/100 | 93% |
| Psicólogo | 380 | 35/100 | 93% |
| Limpieza | 354 | 36/100 | 93% |
| Pintura | 303 | 35/100 | 93% |
| Fisioterapeuta | 363 | 36/100 | 92% |
| Floristería | 309 | 36/100 | 92% |
| Mudanzas | 307 | 36/100 | 92% |
| Electricista | 309 | 36/100 | 92% |
| Arquitectura | 241 | 36/100 | 90% |
| Consultor IT | 366 | 37/100 | 89% |
| Asesoría fiscal | 302 | 36/100 | 88% |
| Panadería | 328 | 36/100 | 88% |
| Cerrajería | 326 | 36/100 | 88% |
| Abogado | 340 | 37/100 | 86% |
| Inmobiliaria | 297 | 37/100 | 85% |
| Farmacia | 297 | 38/100 | 85% |
| Veterinaria | 302 | 38/100 | 82% |
| Óptica | 284 | 38/100 | 80% |
| Restaurante | 303 | 39/100 | 76% |
| Hotel | 269 | 39/100 | 74% |
Geographic breakdown
Differences between cities are minimal: best and worst are separated by less than 5 points
Geographic differences in AI recommendations are smaller than industry differences, but they follow a predictable pattern tied to digital infrastructure density. Madrid and Barcelona — with their larger business ecosystems, higher review volumes, and more websites per capita — show slightly better recommendation rates.
This isn't about business quality. It's about the volume of digital signals available to AI models. A business in a city with 10,000 Google reviews across its sector has a richer data environment than one in a city with 2,000.
The practical implication: businesses in smaller cities need to be more intentional about building their digital footprint, because the local ecosystem provides less ambient signal for AI models to work with.
| City | Businesses | Avg score | Single AI only |
|---|---|---|---|
| Málaga | 1019 | 36/100 | 93% |
| Murcia | 993 | 36/100 | 93% |
| Sevilla | 977 | 36/100 | 92% |
| Palma | 346 | 36/100 | 92% |
| Zaragoza | 952 | 36/100 | 91% |
| Alicante | 948 | 36/100 | 91% |
| Barcelona | 940 | 37/100 | 90% |
| Valencia | 992 | 36/100 | 90% |
| Bilbao | 887 | 36/100 | 90% |
| Madrid | 911 | 36/100 | 89% |
How the 4 AI models compare
Gemini recommends the highest share of local businesses (30%)
Each AI model has a distinct recommendation profile. Gemini and ChatGPT surface the most businesses (30% and 29% respectively), while Perplexity and Claude are more selective at 26%. But the most striking finding is overlap: only 9% of businesses appear in all four models simultaneously.
This fragmentation exists because each AI draws from different data pipelines. ChatGPT leans on Bing's index. Gemini leverages Google's knowledge graph. Perplexity runs real-time web searches. Claude relies on Brave search results and training data.
The strategic takeaway is straightforward: optimizing for a single AI model is a losing strategy. Broad digital presence — structured website data, directory listings, reviews across platforms, authoritative content — is what drives consistent recommendations across all four.
| AI Model | Businesses recommended | Recommendation rate |
|---|---|---|
| Gemini | 3,001 | 30% |
| ChatGPT | 2,896 | 29% |
| Perplexity | 2,557 | 26% |
| Claude | 2,561 | 26% |
The playbook: what recommended businesses have in common
When we analyzed the businesses that do get recommended by 2, 3, or all 4 AI models, a clear pattern emerged. None of them rely on a single silver bullet. Instead, they consistently execute across several digital presence fundamentals:
Implementa schema markup en tu web
Los negocios con datos estructurados (schema.org) tienen más probabilidad de ser citados por las IAs, ya que facilitan la extracción de información clave como horarios, ubicación y servicios.
Aparece en múltiples plataformas y directorios
Las IAs cruzan información de varias fuentes. Si tu negocio solo está en Google Maps, estás limitando tu visibilidad. Registra tu negocio en directorios sectoriales, Yelp, TripAdvisor, Páginas Amarillas y fuentes relevantes para tu sector.
Crea contenido que demuestre autoridad
Las IAs priorizan negocios que tienen contenido original y relevante. Un blog con artículos sobre tu sector, casos de éxito o guías prácticas aumenta la probabilidad de ser citado.
Acumula reseñas en múltiples plataformas
Las reseñas son una señal de confianza para las IAs. Los negocios con más reseñas positivas y diversificadas tienden a aparecer más en las respuestas.
Asegura que tu web es accesible para las IAs
Revisa tu robots.txt y asegúrate de no bloquear a los crawlers de IA. Considera añadir un archivo llms.txt que facilite la indexación por parte de los modelos de lenguaje.
Optimiza tu presencia local
Cuando alguien pregunta a una IA por un negocio en una ciudad concreta, la IA busca señales de relevancia local. NAP consistente (nombre, dirección, teléfono), perfil de Google Business completo y menciones en medios locales marcan la diferencia.
What the data means — and what to do about it
1AI search is a massive blind spot for most businesses
91% of the 9,865 businesses in our dataset are recommended by just 1 of 4 AI models. Three out of four AI-powered search experiences will never surface them. This isn't a future problem — it's happening now, with every AI query a potential customer makes.
2The window of opportunity is closing
AI-assisted search is growing 20-40% quarter over quarter (Statista, SimilarWeb). Google now shows AI-generated answers in 45% of searches. Businesses that build AI visibility today will compound their advantage as the channel grows. Those that wait will find it increasingly expensive to catch up.
3No single AI model covers the full market
ChatGPT, Gemini, Perplexity, and Claude each recommend different businesses. Only 9% appear in all four. A customer choosing Perplexity over ChatGPT may see a completely different set of recommendations. Multi-platform digital presence is the only way to cover the full AI search landscape.
4Service businesses face the steepest uphill climb
Industries like auto repair, photography, and hair salons show single-AI rates above 95%. These businesses typically have thin digital footprints — few directory listings, limited reviews, no structured data. They need the most help, and have the most to gain from AI visibility optimization.
5Start by measuring where you stand
Run Surfeo's free AI visibility scan to see exactly which AI models recommend your business, what they say about you, and where the gaps are. The scan takes 30 seconds and gives you a concrete starting point. Early movers in AI search are already building compounding advantages — the data sources AI models use today will shape their recommendations tomorrow.
Common questions about AI recommendations
Do AI chatbots actually recommend local businesses?
Yes, but selectively. Our data shows ChatGPT mentions 29% of businesses when asked, Gemini 30%, and Perplexity and Claude around 26%. The majority of local businesses are never mentioned at all.
Which industries are AI models worst at recommending?
Service businesses like auto repair (98% single-AI), photographers (98%), and hairdressers (97%) have the largest recommendation gaps. Hospitality (hotels 74%, restaurants 76%) performs significantly better due to aggregator platform presence.
Does business location affect AI recommendations?
Moderately. Businesses in larger cities like Madrid (89% single-AI) and Barcelona (90%) have a slight edge over smaller cities like Málaga and Murcia (93%). The gap correlates with digital ecosystem density, not business quality.
What makes a business more likely to be recommended by AI?
Multi-source digital presence is the strongest predictor. Businesses appearing in directories, review platforms, and with structured website data (schema markup) are recommended 3-4x more often than those with only a basic website.
How does this AI recommendation research work?
We query ChatGPT, Gemini, Perplexity, and Claude with natural-language prompts about businesses in specific industries and cities. Each AI response is parsed to identify which businesses are mentioned, how they're described, and how consistently they appear across models.
Dig deeper into the data
Where does your business stand?
This study was built with the same AI audit technology behind Surfeo. Run a free scan to see which AI models recommend you.
Run your free AI scanAI Recommendation Gap Report by Surfeo
Get the full report (PDF)
We'll send the complete dataset with all charts and sector breakdowns to your inbox.
About this research
Surfeo helps businesses measure and improve how they appear in AI-powered search. We audit your presence across ChatGPT, Gemini, Perplexity, and Claude — the four models shaping how customers discover businesses.
AI Recommendation Gap Report by Surfeo · April 2026