How to get ChatGPT, Perplexity, and Gemini to cite your brand

"How do I get ChatGPT to talk about my company?" We hear this question at least five times a week. On Reddit, on LinkedIn, in discovery calls with our prospects.

The good news: it's neither magic nor reserved for large companies. The bad news: it requires specific work, different from everything you've done in SEO until now.

Let's get straight to it.

Step 1 -- Measure your starting point

Before fixing anything, you need to know where you stand. Open ChatGPT (GPT-4), Perplexity, and Google Gemini. Ask them these questions:

• "What are the best [your service] in [your area]?"
• "Recommend a [your profession] for [typical client use case]"
• "What do you think of [your brand]?"
• "Compare [your brand] and [main competitor]"

Note everything. Who gets cited, in what context, with what qualifiers. That's your baseline.

If none of the three know you, don't panic. That's the case for 80% of the SMEs we audit.

Step 2 -- Define your entity

LLMs reason in "entities." An entity is a clearly identifiable concept: a person, a company, a product, a place. For an LLM to cite you, it must first know what you are.

Concretely, your site must contain a page (or section) that answers these questions in simple, direct language:

Who are you? What do you do? For whom? Where? Since when? What sets you apart?

We're not talking about a vague "about" paragraph. We're talking about text that could be read aloud and understood immediately. "[Name] is a Belgian company specializing in [X] for [target] since [year]." It's basic, but the majority of sites we audit don't even have this.

Get people talking about you outside your own site

This is the main lever. And it's the one everyone underestimates.

An LLM builds its "knowledge" of a brand from multiple independent sources. Your own site is just one source among many -- and it's biased by definition. What really counts:

Press and industry media. An article in a respected publication or even a niche industry medium is worth ten pages on your blog. Contact journalists with a concrete angle, not a generic press release.

Specialized directories and lists. Clutch, G2, Sortlist, your professional federation's directories. Every completed profile is an additional source that the LLM can cross-reference with your site.

Forums and communities. Reddit, Quora, LinkedIn groups, industry forums. Answer questions related to your expertise. Not to advertise -- to demonstrate your competence. Recent LLMs actively explore these platforms.

Wikipedia and Wikidata. If your company is sufficiently notable, create a Wikidata entry (not Wikipedia -- Wikidata is more accessible and equally useful for LLMs). It creates a structured identity card that models love.

Step 3 -- Structure your data for machines

Textual content isn't enough. LLMs -- and especially the retrieval systems that feed them -- also read structured data.

Three technical actions, in order of priority:

Schema.org on all your key pages. Organization on your homepage, FAQPage on your FAQs, Article on your blog posts, Service on your service pages. Not new, but most sites implement it poorly or not at all.

An llms.txt file at the root of your site. It's an emerging format -- a text file that summarizes your company specifically for AI crawlers. We cover it in a dedicated article, but know that Perplexity and others already read it.

FAQ pages that answer real questions. Not the questions you'd like to be asked. The questions your clients actually type into ChatGPT. "What is the best [X] for [Y] in Belgium?" -- your FAQ must answer that, explicitly.

Step 4 -- Create "citable" content

An LLM doesn't cite just any content. It cites content it can easily summarize and attribute. That means:

Affirmative sentences, not passive constructions. "AISOS uses a 3-step method to improve AI visibility" is more citable than "AI visibility can be improved through various methodologies."

Numbers and concrete results. "+340% AI mentions in 4 months for a client in the legal sector in Brussels" -- an LLM can pick that up as-is.

Clear definitions. If you've coined a concept or method, define it explicitly. LLMs are machines for reproducing definitions.

Step 5 -- Repeat and measure

This isn't a one-off project. It's an ongoing process. Every two weeks, re-ask your baseline questions to the three LLMs. Note the changes. Adjust.

The companies that get results are those that treat AI visibility as a standalone discipline -- not as a byproduct of their existing SEO.

One of our clients, an HR consulting firm in Brussels, went from zero mentions to systematic Perplexity citation in 6 weeks. Not with a colossal budget. With 3 press articles, a completed Clutch profile, proper schema.org, and an llms.txt file.

Your turn.