Service / GEO · AI Visibility

Get AI to,
recommend you.

When your client asks ChatGPT, Perplexity or Gemini about your sector, the answer has no page 2. Either you are the source they cite, or you don’t exist in the conversation. GEO works so it’s you.

Search changed form,
not importance.

25%

is how much traditional search volume will drop by 2026, per Gartner, as people ask AI instead of Google. The buying decision now starts in a generated answer, not a list of links.

Reference

Gartner · 2024

What GEO is (and isn’t)

GEO (Generative Engine Optimization) is not keywords: it’s getting generative engines —ChatGPT, Perplexity, Gemini, Google AI Overviews, Copilot— to understand your business, cite you and recommend you. SEO optimized for an algorithm that ranked links. GEO optimizes for a model that writes an opinion. You no longer want first place: you want to be the sentence the AI chooses to write.

What changes

SEO made you findable. GEO makes you recommendable.

The difference

What happens if you don’t work it. And what happens if you do.

Without a GEO strategy

  • ✕AI recommends your competitor and you don’t even appear in the conversation.
  • ✕AI crawlers (GPTBot, ClaudeBot, PerplexityBot) are blocked without you knowing.
  • ✕Your site has no structured data: the model doesn’t understand what you offer or to whom.
  • ✕There’s no way to measure if AI mentions you, ignores you or confuses you with another company.

With a GEO strategy

  • →You’re cited as a reference source: authority lent by the model itself.
  • →Open technical access to AI crawlers (llms.txt, robots.txt): they can read you.
  • →Structured data (Schema.org) that models recognize and cite.
  • →You measure your Citation Rate and AI share of voice every week; you don’t imagine it.

What we work on

It’s not keywords. It’s being the answer.

Five connected levers. Not sold separately: combined by what today stops AI from reading, understanding and citing you.

01AI visibility diagnosis

First we measure what AI says about you today. We don’t assume it.

We audit how ChatGPT, Perplexity, Gemini and Google AI Overviews see you today: where they cite you, where they cite competitors and which questions no one owns yet. We set your baseline Citation Rate and share of voice. Without that baseline, the rest is opinion.

  • −Visibility in ChatGPT, Perplexity, Gemini and AI Overviews
  • −Map of questions where the client is at stake
  • −Benchmark against 3-5 competitors
  • −Baseline Citation Rate and share of voice
02Technical access for AI

If AI crawlers can’t read you, there’s nothing to cite.

We review and open access to AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended): robots.txt, llms.txt publication, speed and rendering. Many portals block AI without knowing and wonder why they don’t appear. This is fixed in days.

  • −robots.txt and llms.txt configured for AI
  • −Crawlers unblocked: GPTBot, ClaudeBot, PerplexityBot
  • −Rendering and speed so they can read you
  • −Report of fixed blocks
03Structured data (Schema.org)

So the model understands what you are and for whom, not guesses.

We implement Schema.org / JSON-LD by content type (Organization, Article, FAQPage, BreadcrumbList, Person, Product). It’s the language AI uses to recognize what you offer, who signs it and why you’re trustworthy. Without structured data, the model improvises, and when it does, it confuses you with another or cites whoever has it.

  • −JSON-LD by page and content type
  • −Organization, author and FAQ schema
  • −Rich Results validation with no errors
  • −Maintenance as models change
04Citable content and brand entity

AI cites who answers with data, not who fills space.

We rewrite key pages with GEO structure: direct answer on top, verifiable data, FAQ and glossary. We reinforce your brand entity (Wikidata, Knowledge Graph) and E-E-A-T signals —verifiable experience, authorship and authority— that a model uses to trust citing you. Content built to be the source of a claim, not to fill the blog.

  • −Pages with direct answer and own data
  • −Brand entity in Wikidata / Knowledge Graph
  • −Author profiles and E-E-A-T signals
  • −Glossary and FAQ with structured data
05Continuous measurement and monitoring

Your AI visibility stops being a guess and becomes a metric.

In GEO the metrics change: no longer rankings and traffic, but whether AI names you, how much and against whom. We measure Citation Rate, AI Share of Voice, Prompt Coverage and AI-referred traffic (trackable in GA4). A dashboard watches the engines weekly and we review it with you monthly with direction, not a junior account.

  • −Citation Rate and AI Share of Voice
  • −Prompt Coverage by target questions
  • −AI-referred traffic in GA4
  • −Monthly dashboard + review with direction

Who is this for?

It’s not for everyone. But if it’s for you, you’re late every month.

GEO makes sense when your client asks an AI before deciding. These are the cases where it shows most.

B2B in shortlists

You fight for AI "shortlists"

When a decision-maker asks "best company for [service] in Spain?", AI gives three names. Either you’re one of them, or you’re not in the conversation.

Product / SaaS

You want to be recommended

Your buyers ask Perplexity or ChatGPT for tool comparisons. If you don’t appear with concrete data, they compare without you.

Knowledge / sector

You want to be the reference

You publish guides, data and judgment. GEO makes AI cite you as a source of definitions and references, not a generic aggregator.

Before creating more content

Know what AI says about you today? We show you in 30 minutes.

Let’s talk →

How we work

Four phases. A deliverable in each.

No endless projects without visible results. Each phase ends with something in your hand you can read, measure and decide on.

01

Diagnosis and baseline

Weeks 1-2. Audit of your AI visibility and the portal’s technical state. Competitor benchmark and opportunity map by query. Deliverable: visibility report + baseline scorecard.

02

Technical implementation

Weeks 3-6. Schema.org, robots.txt for AI crawlers and llms.txt. We fix the blocks that stopped AI reading you. Deliverable: GEO-ready portal, validated schemas.

03

Citable content

Weeks 5-10. Key pages with direct answer, data and FAQ; glossary; entity and authority signals. Deliverable: optimized pages + entity profile.

04

Measurement and improvement

Month 3 onwards. Monthly Citation Rate tracking by engine. We double what works, cut what doesn’t and produce new content. Deliverable: monthly dashboard + review with direction.

How we measure

In GEO the metrics change. These are the ones that matter.

In SEO you saw rankings and traffic. In GEO we measure something else: whether AI names you, how much, against whom and in which questions. No vanity metrics.

Citation Rate

How often AI mentions or cites you when answering about your sector.

ChatGPT · Perplexity · Gemini

AI Share of Voice

Your share of mentions vs competitors in AI answers.

you vs. 3-5 competitors

Prompt Coverage

In what percentage of your sector’s key questions you appear.

target query map

AI-referred traffic

Visits arriving directly from ChatGPT, Perplexity and others to your site.

trackable in GA4

We report it monthly, with your evolution and your competitors’ alongside. A dashboard watches these engines week to week and a monthly review with direction where we decide the next move.

What you won’t find

What we DO NOT do.

In GEO it’s easy to promise and hard to prove. These are the limits that let us work with judgment, not hype.

  • ×

    We don’t promise you’ll appear in every answer — nobody can, and whoever promises it is lying. We sign a method, named deliverables and real month-to-month evolution data.

  • ×

    Nor invent metrics — Citation Rate and share of voice are measured in the engines themselves, not estimated in a pretty PDF.

  • ×

    Nor fill your site with mindless FAQs — citable content is built with own data and senior judgment, not generated filler to trick the model.

  • ×

    Nor lock you in — month to month, full access to your accounts. You stay because it works.

  • ×

    Nor invent case studies — real accounts with real data before you sign anything.

Sectors

Is this for my business?.

It depends on what you want to build.

We work with businesses that no longer want to improvise. That seek direction, not just execution.

EDP
Eleia Energía
Flame Analytics
Fotocasa
Santillana
Viacore
ULMA
Esnova
Ferrovial
Helvetia
ITC
Neodoc

A shared goal:
making every decision count.

Frequently asked questions

What our clients ask us.

Does GEO replace SEO?

No. GEO doesn’t replace SEO: it extends it toward where search is going. Your content has to work in both places at once —classic search and the AI answer— because your clients use both. Much of the technical work (structured data, authority, content with judgment) benefits both. The mistake is choosing one: today you need both.

Do you guarantee I’ll appear in ChatGPT?

No, and be suspicious of anyone who guarantees it. Nobody controls what a model writes in each answer. What we do sign is a method, named deliverables and real month-to-month evolution data. The first measurable results arrive between week 6 and 12 from implementation. And if nothing moves in that window, we tell you first.

How is GEO measured if there are no rankings?

With new metrics. Citation Rate (how often AI cites you), AI Share of Voice (your share of mentions vs competitors), Prompt Coverage (in what percentage of key questions you appear) and AI-referred traffic (visits arriving directly from ChatGPT, Perplexity and others, trackable in GA4). We report it monthly with your evolution and your competitors’ alongside.

What kind of company is GEO for?

When your clients ask an AI before deciding. Three common cases: a B2B company fighting for the "shortlists" AI gives when someone asks "best company for X in Spain?"; a product or SaaS that wants to be recommended when people ask Perplexity for comparisons; and knowledge portals that want to be the cited reference of their sector. If your clients don’t yet use AI to decide, it’s not your moment. If they do, every month without working it is answer share you give away.

Does AI cite at random?

It’s not random, and that’s why it can be worked. A model cites based on your entity (that it knows who you are), your structured data (that it understands what you offer), authority and consensus (that several sources treat you as reference) and whether you cover the real questions people ask. You don’t control the exact answer, but you do control nearly all the signals that tip it your way. That’s GEO.

BLOG

Latest articles.

View all articles →

Let's talk about your business.

The first step is a conversation.

No strings attached, no smoke.

✓We respond within 24 hours
✓No sales pitch. Just strategy.
✓First conversation free, no commitment