What Does AI Say About You?

For years, brands have worried about where they appear on Google. Are we on page one? How well do we rank for the right keywords? Is a competitor above us? How much traffic is coming through search? An entire industry has grown around giving brands the best possible chance of being found when somebody searches.

Now people are searching differently. Ask an AI platform to recommend a hotel for a long weekend in Rome, the best skincare for sensitive skin or a handbag that will survive ten years of daily use, and it can research, compare and assemble an answer for you. The customer may still visit Google, read reviews and eventually land on a brand website, but another voice is beginning to shape what makes the shortlist.

Which raises a new question for brands: when somebody asks AI about you, what does it say?

Search is becoming a conversation

Traditional search taught us to think in keywords. Someone typed “best luxury hotels Rome”, Google returned a page of results, and the battle was to appear somewhere near the top.

AI allows people to be much more specific. Where should I stay in Rome if I want somewhere elegant but don't want to feel surrounded by tourists? Or I'm taking my mother to Italy for her 70th birthday. She loves art and beautiful old buildings. Where should we stay?

These searches tell us quite a bit about the person asking; they contain context, preferences, and clues about what the person values.

To answer, an AI system needs to gather information from different sources, work out what seems relevant and assemble it into something coherent. A brand's own website forms part of that information landscape, alongside editorial coverage, reviews and other sources across the web.

And so, we arrive at GEO.

So, what is GEO?

GEO stands for Generative Engine Optimisation, and broadly describes the work involved in improving how brands and their content are understood and surfaced by generative AI.

It is closely related to SEO, and many of the foundations are shared. Google has been clear that established SEO practices remain important to visibility in its AI search experiences: websites still need to be technically accessible, well-structured and full of accurate, original information. AI systems are drawing on the web, after all, so being understandable on the web remains a sensible place to start.

GEO adds another way of looking at the content itself. If an AI system were trying to understand your brand well enough to recommend it in response to a detailed question, how much would it actually have to work with?

Give the machine something to know

Take a beautiful hotel website. The homepage talks about timeless elegance. The rooms promise refined comfort. The restaurant offers an unforgettable culinary journey, while the destination page invites visitors to discover the city's hidden gems.

It probably looks lovely. It also leaves the hotel undescribed. We wrote recently about luxury's sameness problem, and GEO gives brands another reason to think carefully about the language they use.

The same website becomes much more informative when it tells us the hotel is five minutes from Piazza del Popolo, has 18 connecting rooms, and has a rooftop facing west across Rome. We learn that the restaurant specialises in Roman cooking, that the concierge has relationships with a handful of independent galleries and that the building was once the home of a 19th-century aristocratic family.

Now there is enough detail to answer real questions. Someone looking for connecting rooms can find them. Someone fascinated by Roman food has a reason to choose this hotel. An AI answering a question about hotels with exceptional sunset views has something concrete to draw upon.

The irony is that writing for machines starts to look suspiciously like good writing for humans: specificity, knowledge, context and an actual point of view.

The things that never make it onto the website

A fashion house can tell you why it chose a particular fabric and what it does to the way a jacket hangs. A restaurant knows the farmer who grows its tomatoes and why one dish has survived six menu changes. Ask a good hotel concierge about the city, and they will know which museum becomes unbearable after late morning, or the tiny restaurant still worth crossing town for.

A surprising amount of this knowledge never makes it online. It stays with the people who design the product, cook the food or spend their days speaking with customers. Websites have traditionally focused on the information closest to the sale, while stories about a brand have found their way into PR, newsletters and social media.

GEO creates another reason to bring some of that knowledge into the brand's own content. A very specific question is easier to answer when the details that make a business interesting are there to be found.

Google's current guidance favours original, substantive content created for people, giving the knowledge already sitting inside a business a new value online.

Start with the questions people ask

The questions people put to AI can tell us quite a bit about what matters to them. Someone shopping for a handbag may ask which brands are known for craftsmanship, which bags hold their value, or which designers still manufacture in Italy. A traveller planning a trip to Florence might search for the best hotel in the city, but they might also want connecting rooms, a terrace overlooking the rooftops or easy access to the Uffizi.

Each question gives us a little more context around the decision. Craftsmanship may matter to one person, convenience to another. Someone else may be planning an anniversary and care about what will make the occasion feel special.

We wrote recently about the strange ways we spend money, and how two people can make the same purchase for very different reasons. Questions around a brand offer another way to understand those motivations, especially when considered alongside what you already know from CRM data.

For GEO, those questions also reveal where the content opportunities lie. A brand may already have the expertise to answer them, buried in an existing page or sitting with someone inside the business. Others may point towards stories and information that have never made it online.

What does AI think you are known for?

Brands will naturally want to know whether they appear when somebody asks ChatGPT, Google or another AI platform for recommendations, and which competitors appear alongside them. It is equally worth looking at what they are being recommended for.

A hotel may appear frequently for romantic weekends but barely register for families, despite having excellent connecting rooms and a brilliant children's programme. A beauty brand with years of ingredient research may be known primarily for its packaging and aesthetics. In both cases, the brand is visible, but only part of what makes it interesting is coming through.

Monitoring AI responses can help identify these gaps. You can see which parts of the brand AI already understands well and where there is more to say.

Where Hg2 comes in

At Hg2, we approach GEO as a content discipline. We look at how a brand currently appears across generative search, the questions people are asking around its category and where its existing content gives AI systems enough substance to understand what makes the business relevant.

From there, the work might involve an AI visibility audit, query mapping, restructuring existing content, strengthening key pages or commissioning new editorial where there is a genuine opportunity. The aim is to create a body of content that represents the business accurately and gives it a stronger presence in the conversations happening around its category.

It is closely connected to the work we have always done. Understanding an audience, knowing which questions matter, finding the knowledge inside a business and turning it into compelling content has long been part of good editorial strategy.

The internet needs something to work with

GEO is still relatively new. People are already using AI to search, while brands are still working out how best to understand and measure their presence there.

The brands in the strongest position will probably have something substantive behind them: genuine expertise, clear information and original stories, with enough specificity for both people and machines to understand what makes them relevant.

After years spent trying to get customers to find our websites, we now have to think about the machines helping them decide where to look. And they need something worth finding.