How to Get Mentioned by ChatGPT

There is no switch you can flip to guarantee that ChatGPT mentions your business. Anyone who tells you otherwise is selling fantasy. But there is a practical way to improve your odds substantially: make your website easier for machines to understand, easier to retrieve, and impossible to confuse with a generic competitor.

The businesses that get mentioned by ChatGPT are not always the most prominent or the most linked. They are often the clearest. They have told AI systems exactly what they do, who they serve, and why they are different — in machine-readable terms that do not require inference.

How ChatGPT retrieves recommendations — the actual mechanism

ChatGPT's recommendations come from two different sources depending on how you are using it. In the base model (no web browsing enabled), recommendations come from training data — which means information that was publicly available and indexed before the training cutoff. In web-browsing mode (available in ChatGPT Plus), recommendations can come from live web content retrieved in the moment.

For both modes, the same principle applies: clarity beats volume. A business with a clear, consistent, well-structured web presence is more retrievable than a business with more content but less clarity. Entity graphs — the mental model AI systems use to represent businesses — are built from consistency and specificity, not from word count.

What makes a business 'mentionable' versus forgettable

A mentionable business has a coherent entity profile. The name is consistent everywhere. The service descriptions are specific enough to be matchable to a user intent. The location is clear. The specialisation is not buried in vague positioning copy.

A forgettable business has content that could describe any competitor in its category. 'We help businesses grow with tailored solutions.' That sentence could be the homepage copy for ten thousand firms. AI systems cannot build a distinct entity from it.

In site scans we've run, the most common AI visibility problem is not missing schema — it is service pages that describe the service so generically that a machine cannot distinguish the offer from any other firm in the category.

Organization schema as your entity anchor

The most important single schema implementation for ChatGPT mention probability is a complete, accurate Organization schema block on your homepage. This is your entity anchor — the structured signal that tells AI systems: here is the canonical description of this business, and here is where to find its authoritative web presence.

{
  "@context": "https://schema.org",
  "@type": "LegalService",
  "name": "Northside Legal GmbH",
  "alternateName": "Northside Legal",
  "url": "https://northside-legal.de",
  "logo": "https://northside-legal.de/logo.png",
  "description": "Employment law firm specialising in wrongful dismissal, severance negotiations, and executive employment contracts for employees and employers in Munich and Bavaria.",
  "foundingDate": "2009",
  "numberOfEmployees": {
    "@type": "QuantitativeValue",
    "value": 12
  },
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "Maximilianstraße 35",
    "addressLocality": "Munich",
    "postalCode": "80539",
    "addressCountry": "DE"
  },
  "geo": {
    "@type": "GeoCoordinates",
    "latitude": 48.1374,
    "longitude": 11.5755
  },
  "areaServed": ["Munich", "Bavaria"],
  "sameAs": [
    "https://www.linkedin.com/company/northside-legal-gmbh",
    "https://www.google.com/maps/place/?q=place_id:ChIJXXXXXXXXXXXX"
  ],
  "priceRange": "€€€"
}

Notice what this does: it establishes a specific specialisation (wrongful dismissal, severance negotiations, executive contracts), a specific geography (Munich and Bavaria), and specific sameAs links that allow AI systems to cross-reference this entity against other data sources.

Service schema specificity — the difference between getting mentioned and being skipped

Generic service descriptions are the single biggest reason businesses fail to appear in AI recommendation answers. When a user asks ChatGPT 'which firms handle employment law in Munich?', the model needs to be able to match that query to a specific entity with a specific service profile.

A Service schema block that says 'Employment Law Services' with a description of 'We provide expert legal advice' is unmatchable. A Service schema block that says 'Wrongful Dismissal and Severance Negotiation' with a description that names specific case types, typical outcomes, and client profiles is highly matchable.

{
  "@context": "https://schema.org",
  "@type": "Service",
  "name": "Wrongful Dismissal and Severance Negotiation",
  "serviceType": "Employment Law",
  "provider": {
    "@type": "LegalService",
    "name": "Northside Legal GmbH",
    "url": "https://northside-legal.de"
  },
  "description": "Legal representation for employees facing wrongful dismissal in Germany. We advise on Kündigungsschutzklage proceedings, negotiate severance packages (Abfindung), and handle disputes before the Munich Labour Court (Arbeitsgericht München). Average case resolution time: 4 to 12 months.",
  "areaServed": {
    "@type": "State",
    "name": "Bavaria"
  },
  "url": "https://northside-legal.de/services/kuendigungsschutz"
}

FAQ schema as the direct retrieval layer

FAQ schema is the schema type with the most direct path to AI answer inclusion. When your FAQ answers are specific, authoritative, and structured, AI systems can use them almost verbatim in answer generation. This is not a theory — it is how retrieval-augmented generation systems work.

For ChatGPT mentions specifically, the most valuable FAQ answers are those that address common buyer decision questions: How do I know if I need your service? What is the process? How long does it take? What does it cost? What makes you different from a generalist firm?

What definitely does not work

Several tactics marketed as 'AI SEO' or 'ChatGPT optimisation' have no mechanism for improving mention probability.

• Buying links with the goal of increasing AI training data inclusion — AI systems do not use link graphs for recommendation ranking the way search engines do
• Keyword stuffing service pages with phrases like 'best AI-recommended law firm in Munich' — this is content for search bots, not for buyers, and AI systems recognise it
• Publishing generic thought leadership content that does not describe what you specifically do — AI systems do not reward volume, they reward entity coherence
• Adding schema markup that is not supported by the visible page content — this creates signals that contradict each other and reduce machine trust
• Claiming credentials, specialisations, or locations in schema that are not verifiable elsewhere — consistency across sources is what builds entity trust

5-step implementation plan

1. Audit your homepage Organization schema. If it is missing, thin, or contradicts your visible content, fix this first. It is the entity anchor for everything else.
2. Write specific service descriptions — one per service page. Use concrete language: who the client is, what the service delivers, what the geography is, what the typical process or outcome is.
3. Deploy Service schema on each service page with the specific description you just wrote. Do not aggregate all services in one block.
4. Identify the three to five questions that buyers ask before hiring your firm. Write specific, substantive answers and add them as visible FAQ content with FAQPage markup.
5. Add sameAs links to your Organization schema pointing to your LinkedIn company page, your Google Business Profile, and any other authoritative directory listings.

Loopful supports each of these steps — scan, review, deploy, monitor — without requiring a developer for every change. See how it works. For the specific schema type implementations, Service Schema for Service Business Websites covers the service layer in detail.