Appearing in one AI engine doesn't guarantee automatic presence in another – the same business can be cited by ChatGPT but not appear in a similar response on Gemini. The problem lies in technical factors (crawler blocks), data structure (how each AI extracts information), and knowledge source (LLMs use different data).
The problem nobody is solving
The same brand disappears when you change the language or engine: Galp scores 42.4% in European Portuguese and plummets to 2.9% in Brazilian Portuguese; Tranquilidade had zero mentions in 31 responses in Brazilian Portuguese. But this isn't just a language problem.
Appearing in one AI doesn't guarantee automatic presence in another – Gemini can answer a similar question by prioritising different sources while competitors appear in the response and your company is left out.
What could be happening? Three main reasons.
1. Accidental crawler blocking in robots.txt
This is the most common and easiest error to fix: many companies have robots.txt configured for years with legacy rules, often containing explicit blocks of GPTBot and ClaudeBot that were added "as a precaution" in 2023–2024.
Each AI engine uses a different identifier:
| Engine | Crawler User-agent |
|---|---|
| ChatGPT | GPTBot |
| Claude | Claude-Web |
| Gemini | Google-Extended |
| Perplexity | PerplexityBot |
| Google AI Overviews | googlebot |
What to do: Visit yoursite.com/robots.txt and look for `Disallow` directives targeting any of these crawlers. If they're there, remove them or create a specific rule that only blocks irrelevant pages (such as "thank you" pages, tag archives, or test pages).
Correct example: ``` User-agent: * Disallow: /thank-you/ Disallow: /test/ Allow: /
User-agent: GPTBot Allow: / ```
2. Different data structure between LLMs
When a brand disappears, it's not because the model "decided" to ignore it – it's because there aren't enough signals pointing to it. LLMs learn by association – if the words "Your Brand" and "best tool for [category]" never appear together on authoritative websites, the AI doesn't form a strong vector relationship between them.
Brands that rank well on Google tend to appear more in AI responses because many LLMs use search results as a foundation – maintaining robust technical content, structured data, and authority signals (E-E-A-T) remains fundamental.
But Gemini works differently. Gemini can use publicly available information from Google services, including Google Search and Google Maps. This means that if your brand isn't well positioned on Google Maps or doesn't appear in topics where Gemini relies on Google data, it disappears.
What to do: Audit your Google Maps profile. Check what information appears in Google's Knowledge Panel (search "your brand" and see the sidebar). If it's not there, Gemini will have difficulty citing you.
3. Blocked indexation or content invisible to AI crawlers
Pages that don't pass traditional indexation filters are automatically excluded from AI Overviews – pages must be indexed and eligible to appear in traditional results before being considered for AI features, and indexation issues directly impact a brand's ability to be mentioned.
OpenAI, Anthropic, Meta, ByteDance, and Perplexity don't execute JavaScript during crawling – bots download the JS file but don't run it, which means all content injected into the DOM via JavaScript remains invisible.
If your site is dynamic or relies on JavaScript to load critical content (such as product descriptions, credentials, or data), AI crawlers can't read it.
What to do:
- Use Google Search Console and look for "coverage" – identify pages that aren't indexed.
- Check that critical content (name, description, credentials) is in plain HTML, not JavaScript.
- Test with the URL inspection tool in Search Console – see what the crawler can read.
- If you're using JavaScript for essential content, consider server-side rendering (SSR) or pre-rendering.
Step by step: diagnosing the problem
Before fixing, know what's wrong.
Week 1: Baseline measurement
You can't close a gap you haven't measured – compare how you want AI systems to describe you with how they describe you today by running the same buyer questions in ChatGPT, Gemini, Claude, and Perplexity and recording the results.
Choose 5–10 questions that customers actually ask:
- "What's the best [category] in the UK?"
- "[Your category] with the best value for money"
- "Recommendations for [your category]"
Write the results in a simple table:
| Question | ChatGPT | Gemini | Claude | Google AI Overview | Perplexity |
|---|---|---|---|---|---|
| "Best [X] software" | ✓ Mentions | ✗ Doesn't mention | ✓ Mentions | ✗ Doesn't mention | ✗ Doesn't mention |
| "[Category] in UK" | ✗ Doesn't mention | ✗ Doesn't mention | ✗ Doesn't mention | ✓ Mentions | ✓ Mentions |
Week 2–3: Technical audit
- Robots.txt: Check for crawler blocks (step 1 above).
- Indexation: Use Search Console > Coverage. Look for "Discovered but not indexed" – this means the crawler sees the page but Google doesn't consider it worthy of indexing. LLMs will have the same problem.
- JavaScript accessibility: Use URL inspection in Search Console and compare raw HTML vs. client-side rendering. If there are major differences in important content, you're using JavaScript for critical content.
- Google Maps: If you're local or physical, optimise Google Maps. If B2B or SaaS, check your Knowledge Panel.
Week 4: Prioritisation
With concrete data, identify the pattern:
- If it disappears across all engines: content problem (not indexed, not citeable) or authority issue.
- If it disappears only in Gemini: problem with Google Maps / Knowledge Panel / ranking in Google Search.
- If it disappears in ChatGPT/Claude but appears in Gemini: authority problem or how you communicate (earlier LLMs prioritise Wikipedia, Reddit, articles; Gemini prioritises Google).
What changes after fixing
46.5% of consumers already use tools like ChatGPT and Gemini regularly to research and make purchase decisions. When someone asks an AI which is the best option in your category and the response cites three competitors but not you, the sale started without you in the room.
But it's not just about traffic. AI Overviews reduce organic clicks but maintain brand visibility – for marketing professionals, this shift means rethinking metrics: exposure through AI Overviews functions more like branding than traditional performance.
Measuring traffic alone isn't enough. What matters is knowing whether you're being cited.
- Crawler blocks in robots.txt are the most common error: many companies have legacy rules with explicit blocks of GPTBot, ClaudeBot, or Google-Extended that prevent AI from accessing the site.
- Appearing in one AI engine doesn't guarantee presence in another – Gemini uses public Google data (Maps, Knowledge Panel) while ChatGPT uses diverse web sources, so the problem can be local (Google data) or global (web structure).
- Content injected via JavaScript remains invisible to AI crawlers – OpenAI, Anthropic, and Perplexity don't execute JavaScript during crawling, so dynamic information isn't read.
- The starting point is to compare how you want AI systems to describe you with how they describe you today – run the same questions in ChatGPT, Gemini, Claude, and Perplexity and record the results in a dashboard to identify patterns.
- Brands that rank well on Google tend to appear more in AI responses because many LLMs use search results as a foundation – maintaining robust technical content and authority signals (E-E-A-T) remains fundamental.
Sources
- Medimos como 4 marcas aparecem no ChatGPT e no Gemini
- Erros comuns que fazem sua marca sumir das respostas de IA
- Minha empresa aparece no ChatGPT, mas não no Gemini
- Indexação do Google é chave para aparecer nas AI Overviews
- O ChatGPT é o novo Google
- Google AI Overviews: reduzem cliques mas mantêm visibilidade
- Como otimizar conteúdos para ampliar visibilidade em LLMs e IA
- Melhores Práticas para Otimização de Motores de Resposta
- Cinco estratégias para ganhar visibilidade em IA
- Como aparecer nas buscas por IA: guia prático
See how AI talks about your brand on every topic.