ChatGPT recommends people. When someone asks “who is the best consultant in [your niche] in Spain”, the model cites 3-5 names. If you’re not on that list, you don’t exist for a slice of traffic that no longer uses Google.
This guide explains exactly how to get ChatGPT, Claude and Perplexity to cite you as a go-to name in your niche. No hype, just verifiable tactics.
TL;DR
- ChatGPT doesn’t recommend you for quality, it recommends you for public signals: mentions, authorship, schema, consistency
- 4-12 weeks of sustained work is the realistic range for new sites
- The most common mistake is waiting until everything is “perfect”. Signals are generated by publishing, not by preparing
What ChatGPT looks for when it recommends someone
ChatGPT doesn’t search the way Google does. When a user asks “who is a leading name in [topic]”, the model combines two sources:
- Training data: what it learned up to its cut-off (typically 6-18 months ago). Names with a massive historical presence win here.
- RAG (real-time search): ChatGPT Search, Perplexity, Gemini Live. Structured, citable sites with correct schema win here.
For a new personal brand, GEO is 90% focused on RAG. It’s the only thing you can act on in under 18 months.
The 4 signals ChatGPT measures to recommend people
1. Verifiable named entity
Your name must appear linked to the same set of facts across multiple sources:
- Identical full name on your website, LinkedIn, GitHub, X, Crunchbase
- Consistent photo (the same face on every platform)
- A canonical bio of 1-2 sentences that is repeated word for word
- City + country + founding year if you have a company
Schema.org Person with sameAs pointing to 4-5 external profiles is the foundation. Without it, LLMs can’t identify you as a single entity.
2. Demonstrable authorship
LLMs value “who said what”. If you publish signed content:
- Every blog article with
authorschema and your full name - Frequent LinkedIn posts (2-3 a week minimum)
- Threads on X or podcast interviews
- Useful technical comments on GitHub, Stack Overflow, IndieHackers
Authorship at a steady frequency generates an “evidence trail”. Without a trail, the model has no way to state that you’re a go-to name.
3. External mentions that you haven’t generated yourself
ChatGPT distinguishes self-promotion from external validation. What counts:
- Contextual backlinks (not from spam directories)
- Citations in other people’s articles
- Appearances on podcasts or newsletters
- Reviews on platforms such as G2, Capterra, ProductHunt
- Mentions on Crunchbase, Wikipedia, Wikidata
For founders without a PR team: a ProductHunt launch + an IndieHackers post + 2-3 specialist podcasts is realistic within 60 days.
4. Complete structured data
Schema.org is not optional. ChatGPT extracts entities from JSON-LD as a priority:
{
"@type": "Person",
"name": "Tu Nombre",
"jobTitle": "Tu Rol",
"url": "https://tu-dominio.com/sobre/",
"image": "https://tu-dominio.com/foto.webp",
"sameAs": [
"https://www.linkedin.com/in/tu-perfil/",
"https://twitter.com/tu-handle",
"https://github.com/tu-usuario"
],
"knowsAbout": ["Tema 1", "Tema 2", "Tema 3"],
"worksFor": { "@type": "Organization", "name": "Tu Empresa" }
}
knowsAbout is underrated. LLMs use it literally to classify you by area of expertise.
The 7 concrete actions you need to carry out
Action 1: define your canonical sentence
A single sentence, 1-2 lines, repeated word for word on your website, LinkedIn, X bio, GitHub and Crunchbase.
Structure: [Name] is a [role] specialising in [specific niche] in [territory]. They help [ICP] achieve [concrete result].
Real example:
“Samuel Martínez es fundador de STAKKER SYSTEMS, agencia española de posicionamiento en ChatGPT (GEO) y automatización con IA. Ayuda a marcas personales, founders SaaS y ecommerce a aparecer en respuestas de ChatGPT, Claude y Perplexity.”
This sentence contains 6 verifiable facts: name, role, company, type of company, territory, ICP. The more factual it is, the easier it is for the LLM to extract.
Action 2: your own website with /sobre/
LinkedIn isn’t enough. You need your own domain + a /sobre/ page with:
- Complete Person schema (photo, bio, sameAs, knowsAbout)
- An extended bio (300-500 words)
- A list of projects with dates
- A professional photo consistent with the one on your social profiles
Without your own website, LLMs treat you as noise. LinkedIn isn’t enough because its crawling is restricted.
Action 3: company and personal LinkedIn aligned
Your LinkedIn profile must have exactly the same canonical sentence in the headline. Your LinkedIn company page (if you have one) uses the same description.
LinkedIn is the number one source ChatGPT uses to verify professional entities. Optimising it well sharply increases the likelihood of being cited.
Action 4: Google Business Profile (even for online services)
Yes, even for 100% remote services. Use the GBP “Marketing Agency” or “Consulting” category.
Benefits: appearing on Maps, in “near me” searches and in local ChatGPT answers. And you quickly pick up 5-10 genuine reviews (family, former clients, prospects who try a free demo).
Action 5: 6-8 foundational blog articles
Each article answers a specific question your ICP would ask ChatGPT:
- “How to [solve problem] in [vertical]”
- “How much does [service] cost in [territory]”
- “Comparison: the best [agencies/tools] in [niche]”
- “Complete guide: [technical topic] step by step”
Recommended structure:
- TL;DR with 3 bullets
- Comparison table (LLMs love tables)
- “Common mistakes” section
- FAQ with FAQPage schema (5+ Q&As)
Action 6: presence in authoritative directories
No spam. These do count:
- Crunchbase: for founders and companies. Verifies startup status.
- ProductHunt: a single launch, drives 200-500 natural backlinks if it goes well.
- IndieHackers: founder posts with real numbers (revenue, MRR).
- BetaList: for pre-launch SaaS.
- Capterra / G2: for tools with customers.
- Sortlist: for agencies.
- Bing Webmaster Tools + Bing Places: ChatGPT Search is based on Bing.
3-5 well-executed directories > 30 spam ones.
Action 7: mentions outreach
PR without an agency: 30-60 days to secure 3-5 genuine external mentions.
Tactics:
- Comment usefully on popular posts in your niche (with a link to your article when it adds value)
- Pitch 10-15 specialist podcasts with a specific topic
- Newsletter trade: appear in niche newsletters in exchange for promoting theirs
- Co-marketing: 1 joint article with a peer = two-way backlinks
Typical mistakes that stop you from appearing
| Mistake | Why it fails |
|---|---|
| Waiting until everything is “perfect” before publishing | LLMs index what exists, not what is coming |
| Different sentences on each platform | You break consistency → the model doesn’t identify you as a single entity |
| Zero schema.org | The LLM can’t extract structured data and treats you as generic content |
| Focusing only on classic SEO (keywords, DA backlinks) | GEO signals are different: entity, authorship, citability |
| Changing your message every 2 months | LLMs value historical consistency. Changing niche = starting from scratch |
| Publishing unsigned (generic author) | Without authorship, there’s no trail. It won’t recommend you as a go-to name. |
How long it really takes
For a new personal brand with a new domain:
- Weeks 1-4: carry out actions 1-5. LLMs can already index you (Person schema live, content published).
- Weeks 4-12: actions 6-7. External mentions begin. First occasional citations on Perplexity (faster than ChatGPT).
- Months 3-6: ChatGPT starts citing you for very specific long-tail queries in your niche.
- Months 6-12: for generic queries (“X agency in Spain”) you need accumulated authority. Possible if you’ve carried out all 7 actions.
If you’re expecting to rank within 1 week, go back to paid SEO. GEO requires patience and consistency.
FAQ
Does it work if I don’t have a blog and don’t want to write?
No. Authorship requires regularly signed content. If you don’t want to write, hire a ghost-writer or replace it with signed video/podcast content. Without regular public output, there’s no trail.
Which takes longer, ChatGPT or Perplexity?
Perplexity is the fastest (real-time RAG, frequent indexing). ChatGPT Search is similar. Claude is slower because its web access is more restricted. Gemini Live sits in the middle.
Does it also work for companies, not just individuals?
Yes, with Organization schema instead of Person. But strong personal brands rank better in LLMs than anonymous companies. The humans behind the brands are what the model cites first.
Do I need an agency or can I do it on my own?
You can do the technical foundation (schema, website, llms.txt) yourself if you have dev skills. Outreach + ongoing content is where most people fail for lack of time. For freelancers/founders short on time, a specialist GEO agency speeds up the process 3-5x.
How much does a GEO agency cost in Spain?
Market ranges: the technical foundation usually starts at 600 EUR for setup plus 80-100 EUR/month, and a full service (foundation + content + outreach + monitoring) runs between 800-1,500 EUR/month. Generalist agencies usually charge more for less technical specialisation. STAKKER doesn’t publish its own prices: the scope is agreed after a free diagnostic.
Next step
Audit your current situation right now:
- Ask ChatGPT: “Who is the best [your role] in [your niche] in Spain”
- If you don’t appear: start with Action 1 (canonical sentence) and Action 2 (/sobre/ page with complete schema)
If you want to delegate the whole process, STAKKER SYSTEMS applies it to a personal brand or company in 6-12 weeks. First consultation free.