Here’s an experiment.
Open your favorite artificial intelligence (AI) tool and type in the prompt “Show me a list of the top ten [your company type] in [your location].”
Did your brand show up in the answer? Welcome to the world of AI brand visibility.
Search is moving toward AI-powered results at speed. According The State of Marketing, a report from Hubspot, nearly half of marketers (49%) say that web traffic from search has decreased because of AI-generated answers.
Marketing agencies argue that brands need to act (and spend) now. The same firms that have generated billions of dollars from promoting search engine optimization (SEO) services are now rapidly pivoting to what is being called answer engine optimization (AEO) or generative engine optimization (GEO). We’ll stick to GEO.
Sam Weston is head of AI and marketing at 80 DAYS, a full-service creative and digital agency that works with around 550 hotels worldwide.
Weston agrees there is considerable hype around the GEO shift but says it is grounded in strong technical SEO.
“If you've been working on technical SEO for the last 15-20 years, then you've laid some really good groundwork. Where GEO differs is that it's around managing your entire digital profile online and having a really clear ownership of all of the different channels where you're represented online,” Weston said
“We talk about GEO in terms of three Cs: clarity around your positioning and offering, having consistency around the messaging. Credibility is increasingly important. A lot of these large language models (LLMs) are citing TripAdvisor, Conde Nast and Forbes, so having credible signals that what you say as a brand is reinforced by what others say is increasingly important.”
On the technical side of things, Weston recommends making sure your schema markup is solid so that LLMs can accurately understand who you are and what you're offering and surface you for the types of questions that are being asked.
Curacity, a company that generates articles and newsletters for more than 40 media brands, including Travel + Leisure and Conde Nast Traveler, has recently launched an updated distribution platform, VISTA, to help travel brands with their GEO.
Nick Slavin, Curacity’s chief executive officer and co-founder, believes individual hotels will struggle to generate visibility by creating their own written and video content as it is “difficult to scale and requires enormous investment.”
“A few brands, such as Marriott with Marriott Bonvoy Traveler, have built full editorial operations to establish authority in search. But as travelers shift from traditional search to conversational AI, LLMs tend to cite third-party sources rather than brand-produced content,” Slavin said.
“The opportunity for brands isn’t just to be found by AI, but to shape demand by appearing within the authoritative editorial environments that inspire travelers before they’ve decided where to go or where to stay. There’s still a marketing funnel, even if the entire journey takes place on a single AI platform.”
Slavin says the challenge for brands is that with SEO they were much more in control.
“Brands had a fairly clear roadmap: build a well-optimized website and earn backlinks from authoritative sites. Roughly half of the ranking factors were within the brand’s control. AI changes that balance. LLMs rely far more on how brands appear within trusted third-party sources and the broader context around them.”
How much can brands control visibility?
Before blowing the marketing budget on GEO, it’s probably worth looking at a recent experiment carried out by market research firm SparkToro.
The firm asked 600 people to run prompts through ChatGPT, Anthropic’s Claude and Google’s AI Overview or AI Mode. The prompts, which were the same each time, asked for the top brands in a number of categories, including chef’s knives, cancer hospitals and best TikTok accounts about US politics.
SparkToro co-founder Rand Fishkin shared the results here. The research found that there is a less than one in 100 chance that ChatGPT or Google’s AI, if asked a hundred times will give you the same list of brands in any two responses. Claude is just slightly more likely to give you the same list twice in a hundred runs.
Hotelrank.ai, which sells AI visibility services, then extended SparkToro’s idea into the hotel space. The company carried out its own research involving 4,000 hotel queries in Google AI Mode across eight cities and using five different types of prompt, including refinements such as star ratings and personas.
The company’s co-founder Nicolas Sitter said that hotel searches in AI are around 50 times more consistent than the brand categories SparkToro tested
“The number one hotel is stable 50.5 percent of the time when you run the same query,” he wrote on LinkedIn.
Sitter postulates that the reason for the difference between brands and hotels is that “hotels are structurally constrained in ways brand queries aren’t.”
For example, requesting a Paris hotel can only return a Paris hotel. “Geography is a hard filter,” he wrote.
Sitter’s takeaway from all this is that it is “maybe worth a bit of optimization.
Mike Coletta, senior manager for research and innovation at Phocuswright, said, “While LLM visibility appears to be a bit arbitrary and no guarantees can be made, the goal of LLMs is the same as it was for general search: to provide the best answer to the query.”
“Both platforms can only rely on digital information to accomplish this, so it still stands to reason that a company’s digital presence is paramount. This means lots of positive mentions, which means offering services people love and doing all the normal marketing tasks to get the word out, encourage travelers to leave reviews, which can be done at any scale.”
Adam Harris, CEO of cloud-based property management system Cloudbeds, is not convinced of the value of GEO services.
“AI visibility does not exist. It's total bullshit,” he told PhocusWire. “If anyone actually believes that they can push you higher in the rankings, they're selling you snake oil. The things that are going to help you lift are the same things that help you rank in traditional search: good quality service, great location, positive reviews.”
Cloudbeds has carried out its own research on visibility using hundreds of queries across the major LLMs.
According to Harris, “Every AI-recommended property had an average of 8.6 out of 10 score or higher. For the top results, each point of addition represented 11% higher discoverability. Thirteen percent of AI citations came from above-average, fresh content on the website, conversational language and really good metadata around the address.”
The AI signals provided by big online travel agencies (OTAs), such as Booking.com, cannot be ignored either, he said.
“Fifty-five percent of the AI citations came from OTAs, so if you're not professionally listed across OTAs that are spending 20 plus billion dollars a year trying to feed their content out there, you lose some of the distribution.
“No agency required,” said Harris definitively. “If hotels are already doing those things, they should be really, really good.”