A GEO audit measures how a brand is understood, cited, and recommended by generative artificial intelligence systems such as ChatGPT, Gemini, Perplexity, or Claude. To be meaningful, it must do more than test keywords: it must reproduce customers’ real-life situations, intentions, and language.

What Is a GEO Audit?

When a customer asks an AI to recommend a company, they do not phrase their request as they would on Google. They describe a situation, specify their constraints, sometimes express a pain point, and expect an answer they can act on directly.

This shift is changing how brands need to think about visibility. It is no longer enough to rank highly in a list of results: brands must be understood, cited, and recommended by AI in the situations that truly matter to customers.

Generative Engine Optimization, or GEO, refers to the practices used to improve a brand’s visibility in AI-generated responses. The GEO audit is the starting point: it measures where, how, and to whom the brand appears. But to measure its presence accurately, the right questions must be asked of the right models. This is where customer experience expertise becomes decisive Generative Engine Optimization. 

GEO and SEO Audits: Why GEO Is Not Simply SEO Repackaged

GEO and SEO are closely related. A well-structured website, high-quality content, and strong editorial authority remain essential for being recognized as a credible source. GEO therefore does not replace SEO; it builds on its objectives. But usage patterns differ.

On a traditional search engine, a user might enter a few keywords such as “pallet transport for SMEs.” When asking an AI, the request becomes conversational:

“Which transport company would you recommend to an industrial SME looking to reduce delivery delays without making its operations more complex?”

The query is no longer just about a service offering. It incorporates a type of company, a context, an objective, a pain point, and selection criteria.

An approach focused on search volumes or keyword variations therefore measures only part of the picture. It shows whether a brand appears for certain queries, but not whether it is recommended at the most important decision-making moments for its customers.

Start with Experiential Profiles, Not Just Keywords

Before writing a single prompt, we define the customers the company wants to reach. The goal is not simply to create personas based on age, job title, or company size, but to build experiential profiles: people in a concrete situation, with a particular level of knowledge, constraints, objectives, and way of expressing their needs.

Client Case Study: A GEO Audit for a Global Logistics Company

During a recent GEO audit for a global logistics company, we worked with three profiles:

•      An SME owner looking for a simple, reassuring solution, without a dedicated logistics team;

•      A logistics director at a mid-sized industrial company, familiar with industry terminology and responsible for tracking specific performance indicators;

•     A garage owner with an occasional, urgent shipping need, under pressure from their own customer.

These three people are looking for a transport solution, but they do not ask an AI the same questions. The SME owner wants to reduce the stress caused by urgent situations. The logistics director focuses on OTIF, SLAs, and total cost. The garage owner is mainly looking to minimize vehicle downtime.

A generic question such as “What is the best B2B carrier?” would erase these differences. Yet these are precisely the differences that determine whether a recommendation is relevant.

How Should You Choose Which LLMs to Audit?

Defining the right profiles also helps determine which environments to assess for visibility. Users do not all have the same digital habits, level of familiarity with AI, or professional tools. A garage owner searching quickly on a phone will not use AI in the same way as a journalist, consultant, or logistics director.

Some profiles will mainly use ChatGPT as a general-purpose assistant. Others will turn to Gemini within the Google ecosystem, Perplexity to find recent sources, Claude to analyze documents, or an AI integrated into their work environment. The choice depends in particular on:

•      The occupation and level of expertise;

•      Search habits;

•      The tools provided by the company;

•      The device and ecosystem being used;

•      The country, language, and usage context.

However, avoid automatically assigning an LLM to a profession: these assumptions must be checked against data, surveys, or customer interviews. Strong visibility on an LLM rarely used by a priority audience does not make up for being absent from the tool they rely on. The audit must therefore connect the customer profile, the situation described, and the LLM they are actually likely to use.

Put Prompts Back into the Customer Journey

The same profile will not make the same requests throughout their journey. At first, they seek to understand their problem. They then compare solutions, assess risks, look for evidence, and may eventually ask for an explicit recommendation. A GEO audit must cover these different intentions:

•      Understand a need or solve a pain point;

•      Discover the available solutions;

•      Compare offers or providers;

•      Assess risks and benefits;

•      Find recognized players in the market;

•      Select a solution.

This mapping reveals the areas in which a brand can become a reference even before users directly search for a provider. It follows the same intent-based logic that underpins effective marketing orchestration. Prompts should also vary in expertise level, business context, urgency, selection criteria, and the rational or emotional factors driving the decision.

In our logistics audit, 90 prompts were developed from three profiles. An analysis of their semantic branches—the query fan-out described by Google for its AI Mode—made it possible to study 972 queries across ChatGPT, Gemini, and Perplexity query fan-out described by Google for it's AI model. 

But volume only matters if the starting point is right. Automating hundreds of prompts that are poorly aligned with customers does not make an audit more relevant; it simply industrializes a flawed brief.

Our Six-Step GEO Audit Methodology

With T&S Digital Generated, our GEO audit methodology is structured around six steps:

1.     Define strategic priorities: specify the company’s offerings, target audiences, markets, and objectives.

2.     Build experiential profiles: identify their situations, needs, pain points, and decision criteria.

3.     Map the customer journey: cover discovery, comparison, reassurance, and selection intentions.

4.     Create realistic prompts: vary the language, context, constraints, and level of expertise.

5.     Measure and interpret visibility: analyze presence, position, persistence, competitors, and the sources used.

6.     Weight results according to usage: put visibility on each LLM into perspective with the habits of priority profiles.

This approach moves beyond a general visibility measure to a strategic diagnosis: Which profiles is the brand recommended to? In which situations is it absent? For which benefits is it recognized? Which players control the narrative?

Measure Visibility in AI—and Its Relevance, Too

An overall average can suggest satisfactory visibility while concealing significant gaps. A brand may be cited frequently for expert users but be absent from responses intended for non-experts. It may appear when a customer is looking for technical performance, but not when they prioritize simplicity, support, or reassurance.

These gaps point less to a ranking problem than to a mismatch between the brand’s narrative and the expectations of its different customers.

Customer experience expertise turns these findings into operational decisions. When visibility is lacking, simply producing more content is not enough. It is necessary to determine which profile to address, in which situation, at what stage of the journey, using what level of language, and supported by what evidence.

Recommendations may include guides, FAQs, comparison pages, use cases, infographics, or videos. They may also involve simplifying existing content or strengthening the brand’s presence in the media, comparison sites, and industry organizations that AI systems use as sources.

SEO and Customer Experience: Two complementary areas of expertise

SEO agencies have essential skills for analyzing a website’s structure, authority, and ability to be interpreted by search engines. These skills remain vital to a GEO strategy.

Customer experience expertise adds another dimension: understanding why, when, and how a person queries an AI, then checking whether the response aligns with the brand’s objectives. This is at the heart of Acemis, a consulting firm specializing in data and AI for customer experience.

Starting with search engines leads to the question: “For which queries does the brand appear?” Starting with customers leads to a more strategic question:

“In which situations do we want to be understood, cited, and recommended—and to which customers?”

It is this second question that makes GEO not merely another technical visibility exercise, but a genuine driver of differentiation and performance.

FAQ: GEO Audit

What is a GEO audit?

A GEO audit (Generative Engine Optimization) analyzes how a brand is understood, cited, and recommended by generative AI systems such as ChatGPT, Gemini, Perplexity, or Claude. It measures the brand’s presence and position, the competitors mentioned, and the sources used by the models.

What is the difference between a GEO audit and an SEO audit?

An SEO audit assesses a website’s position in search engine results based on keywords. A GEO audit assesses a brand’s presence in generated responses to conversational queries that incorporate context, constraints, and selection criteria. The two are complementary.

How many prompts are needed for a reliable GEO audit?

The number matters less than how closely the prompts align with real customers. For example, our logistics audit used 90 prompts based on three profiles, resulting in 972 queries analyzed after their branches were taken into account.

Which LLMs should be included in a GEO audit?

Those your priority customers actually use. ChatGPT, Gemini, and Perplexity cover a large share of usage, but the choice should be validated through data, surveys, or customer interviews.

Make GEO a Tool for Customer Strategy

A brand that is visible in AI is not necessarily well positioned. It becomes well positioned when it appears at the right time, for the right profile, and for the right reasons.

With T&S Digital Generated, we have developed a GEO approach that combines digital expertise, mastery of AI environments, and in-depth knowledge of customer journeys and behavior. From defining profiles to measuring visibility, we help companies make GEO a tool in service of their strategy and integrate it into the orchestration of their marketing activations. 

Would you like to assess how your brand is understood, cited, and recommended by AI? Let’s discuss your challenges and work together to build a GEO audit tailored to your customers, offerings, and priorities.

By Gauthier Dessales-Quentin, Manager at Acemis, expert in digital strategy, SEO, and GEO.