Help us prepare for our appointment
What does LLM referencing involve? It’s about making your brand visible, citable and recommendable by language models – in short, a reference point for AI. More specifically, when users ask ChatGPT, Gemini, Claude or Perplexity directly: ‘What’s the best solution for…?’ “Which company should I choose for…?” or even “Compare the top players in the market for me”, your brand or company needs to be part of the answer. So, how do you become a go-to source for LLMs? Which strategies should you focus on – SEO, GEO, authority or citations? How can you measure your visibility? And, above all, which agency should you choose for your LLM search engine optimisation? We’ll explain how to come out on top and make your brand a go-to source, for both internet users and AI systems.
Key takeaways:
In summary, LLM search optimisation is not about replacing SEO, but about expanding your visibility strategy to include new conversational search pathways. The aim is to make your brand a source that is sufficiently relevant and credible to be naturally associated with its area of expertise by search engines and AI systems.
LLM stands for Large Language Model. It’s the engine that powers ChatGPT, Claude, Gemini and Perplexity. Essentially, it’s a system that has read an enormous amount of text and, based on that, generates a response word by word, guessing the most coherent next sequence each time. ChatGPT, Gemini and Claude are the best-known examples of this.
Featuring your brand on these models has nothing to do with how Google works. A search engine simply returns a list of links. An LLM, on the other hand, processes the information and generates a completely new text, using its own words. Your content can feed into its response… without the reader ever knowing you’re the source 🫥 Too bad.
There are two ways for a model to retrieve information: either from what it has learnt during training, or in real time via a system that fetches up-to-date data from the web (the well-known RAG). In both cases, your objective remains the same: to become the reference source that the model chooses to draw upon.
It is no longer simply a matter of ranking a page in search results, but also of becoming a source that is sufficiently relevant, identifiable and credible to be referenced or cited in an AI-generated response.
This is where GEO (Generative Engine Optimisation), or LLM SEO, comes into play.
💡 Good to know: LLM search optimisation is no longer limited to ChatGPT, Gemini or Perplexity. Google is increasingly integrating generative AI directly into its search engine, notably through AI Overviews and its AI Mode. This means users can get a summary of the answer even before clicking on the traditional blue links. This development brings SEO and GEO even closer together.
Traditional SEO focuses on keywords, link building and well-structured headings. It remains useful, as many LLMs still rely on pages indexed in the traditional way. But the game has taken on a whole new dimension.
On Google, a good ranking aims to generate a click. With an LLM, your brand can become part of the answer itself, sometimes accompanied by a link to your website. The challenge is therefore changing: your content must not only be visible, but also clear, reliable and relevant enough to be included or cited by the model rather than by a competitor.
Three game-changers:
And that’s not all: each AI sources information in its own way, which only serves to make things even more complicated ^^. Content tailored for Google’s AI Overviews offers no guarantees on Perplexity, which sources information from different places and in different ways. Ad’s up breaks down these nuances search engine by search engine in its analysis of the launch of AI Overviews and AI Mode in France.
💡 Good to know: according to a study published in 2026, the appearance of an AI Overview in Google causes organic outbound clicks to fall by nearly 40 per cent on average. This really underlines the urgency of focusing on LLM visibility, and not just traditional SEO.
Google Search Console already allows you to view certain data relating to visibility in AI-enhanced results, particularly via impressions 🤷 But to gauge your brand’s presence more broadly in the responses generated by various LLMs, there are two simple steps you can take to gain a clearer picture. The first, which is simple and accessible: ask your strategic queries directly to ChatGPT, Perplexity or Gemini. Does your brand appear? In what context? Is it cited as a source? And, most importantly, which competitors feature in its place? The second, more in-depth approach: specialised tracking tools allow you to automate the monitoring of your brand’s citations, mentions and rankings within AI responses. A good LLM SEO agency should be able to structure this monitoring and analyse this data to identify opportunities for optimisation.
| Traditional SEO | LLM Indexing | |
| Objective | Rank in positions 1–3 | To be mentioned in the generated response |
| Key indicator | Ranking, organic traffic | Citation rate in AI responses |
| A winning format | Long-form articles, keywords | Standalone sections, structured data |
| Sign of trust | Backlinks, domain authority | E-E-A-T, multi-source consistency |
| Rate of change | A few major updates each year | Ongoing, almost monthly |
Because adding three FAQs to your articles and an llms.txt file doesn’t constitute a GEO strategy. 😉
A truly specialised agency must be able to work on several areas simultaneously: technical SEO, editorial strategy, structured data, brand authority, presence on third-party sources, and monitoring visibility in generative environments.
The right approach, therefore, is to look for an agency capable of combining SEO, content, data and brand awareness, rather than a service provider offering a ‘miracle solution’ to ‘rank on ChatGPT in 30 days’.
The virtuous cycle, in 7 steps:
💡 Advice from our experts:
Ask how your LLM visibility is measured
👉 A reputable agency should be able to explain which prompts are tracked, which models are analysed and how it measures changes in your citations and mentions.
Be wary of promises of quick
results
👉 There is no magic formula for getting cited by ChatGPT or Gemini in a matter of weeks. Opt for a methodology based on SEO, authority, expert content and analysis of results.
Check that the agency works beyond your website
👉 LLM visibility also relies on signals from third-party sources. A good strategy must therefore take into account the press, partners, reviews, forums and, more broadly, the brand’s ecosystem.
Ask for concrete
evidence
👉 Case studies, examples of mentions, changes in visibility or prompt analyses: an expert agency must be able to illustrate its method with results and real-world examples.
Choose an agency that adapts its approach
👉 Generative models are evolving rapidly. Your partner must actively monitor developments, test new tools and regularly adjust the strategy in line with changes to LLMs.
At Ad’s up, the issue isn’t about choosing between traditional SEO and LLM search optimisation.
The aim is rather to develop a strategy that can deliver results on search engines today, whilst also increasing the brand’s visibility within new conversational pathways.
This approach involves, in particular, analysing the brand’s presence in generative responses, identifying high-potential topics and prompts, optimising the website from both a technical and editorial perspective, and developing consistent authority signals.
The idea? To move beyond simply aiming to rank first for a search query, and instead become a brand that search engines, internet users and AI naturally associate with your expertise.
Would you like to find out where you stand when it comes to conversational AI? The Ad’s up team can review the situation with you and lay the groundwork for a roadmap.
It refers to all the actions that make a brand sufficiently clear, credible and well-documented for a language model to mention, cite or recommend it in its responses, without relying on a traditional list of links.
The terminology varies depending on who’s talking, but the essence remains the same: ensuring that a generative AI selects your brand as a reliable source. GEO remains the most commonly used term, whilst ‘LLM SEO’ is its French equivalent.
No, not automatically. It helps, because many models rely on the traditional index, but an LLM also takes other signals into account, such as the consistency of the information and the authority established elsewhere on the web.
There’s no guaranteed timeframe. Unlike Google updates, which occur every few months, these models evolve continuously, which can speed up or slow down the effects depending on your sector.
Ask them to show you how they measure LLM visibility in practical terms, using which prompts, and on which models. A vague or theoretical answer speaks volumes.
LLM search optimisation is a discipline in its own right, with its own methodology, its own set of rules, and a pace of development that never slows down. Brands that make the switch now, rather than waiting for everyone else to jump on the bandwagon, will have a massive head start in the game. 🚀
Ready to find out whether your brand is already part of the answer, or whether it’s still in the AI’s blind spot? The Ad’s up team can check this for you and help you plan the next steps.
