Last month, California designer Shaun Crha got a little glimpse of the future. A potential client called him out of the blue, and as it turned out, the lead came from AI. “She has a Spanish revival home and wants to preserve it, but for it not to feel like a museum,” says the founder of Long Beach–based Wrensted Interiors. “She was talking about it with ChatGPT and my name came up.”

Crha isn’t the first designer to get a lead from an AI search. What started as a novelty a few years ago has become a steady drip-drip in many a firm’s pipeline. It’s not a deluge, and word-of-mouth still reigns supreme, but chatbots like Claude, Perplexity and ChatGPT are starting to connect clients with designers.

So, when a homeowner fires up their bot of choice and starts asking about design, how do you make sure your name comes up? “If that’s where people are searching for information, it does make me question: ‘How do I make sure I stay visible?’” says Crha.

Online search has changed so quickly and completely, soon we may barely remember the era when “Googling” meant pecking through lists of blue links on a hunt for the crucial nugget of information. Now many industry estimates say that 50 percent of Google search results generate an AI-powered answer, while search through chatbot queries continues to slowly chip away at the edges of Google’s dominance. All of the momentum points in a single direction: Looking for something online often means using AI.

Changes to what shows up when you type “best BBQ Houston” into a search bar may not seem like the biggest deal in the world. In fact, it’s an enormous disruption to the state of play on the internet. Billions of dollars of commerce depends on search results, and the retooling of the technology to run on AI is the most significant change to the online ecosystem since, well, Google itself.

With that change comes opportunity, money, and a new discipline: GEO, or “generative engine optimization.” In fact, the field is so new that there’s no consensus yet on a name—though AEO (answer engine optimization) and AIO (artificial intelligence optimization) may differ slightly from GEO, the three terms are used almost interchangeably. One thing everyone can agree on is that there’s plenty of money in helping companies grow their presence in AI searches. An AI visibility startup named Profound raised more than $150 million in its first 18 months. It’s one of dozens.

The real opportunity may be on the other side of the equation. The shift to AI-driven search has reshuffled the deck, opening up a window for companies (and designers) who were never able to crack the SEO code and rank at the top of Google searches. There’s a sense—accurate or not—that now is the time to get in on GEO, before AI-driven search becomes either cynically gamed or contaminated by ads. “SEO has gotten to a point where it’s pay-to-play,” says Byron Cordero, a PR consultant who has recently delved into AI visibility services. “I don’t want to say it’s a monopoly, but for smaller brands that can’t afford [to spend] $3,000 a day for Google Ads, AEO is great because you’re able to still compete.”

The nature of AI-driven search only heightens the stakes. Whereas “classic” Google searches might have given users a page of 10 links to click on, chat-based searches tend to only surface two or three options. If you’re not one of them, you’re all but invisible.

To illustrate the stakes, digital PR firm 5W collaborated with Miami-based design magazine Haute Living’s designer directory to produce a report called The Designer AI Visibility Index. The methods were simple: 5W fed prompts into ChatGPT, Claude, Perplexity and Google’s AI search over the course of a month, asking each to recommend design pros for a fictional $50 million oceanfront estate project in four markets. The company then tracked “citations”—or how often the chatbots recommended which designers.

Unsurprisingly, the really big names like Kelly Wearstler, Martyn Lawrence Bullard and Atelier AM all showed up a lot. But according to the report, there was an interesting division a few tiers below—up-and-coming firms that had developed a robust online presence were outranking AD100 and Elle Decor A-List designers who were quieter on the Internet. Many didn’t show up in 5W’s study at all.

This isn’t exactly rocket science—showing up online puts you into online searches. But it gets the point across effectively: In a world driven by AI-powered search, it’s easy to disappear. “If you’re not speaking to the machines in a way that they can hear you and they understand you, good luck,” says 5W founder Ronn Torossian.

The challenge? GEO, to put it kindly, is murky. The technology is very new and changing every month. The AI engines are black boxes—even the companies developing them don’t fully understand what’s happening underneath the hood. And the gold-rush energy around AI is leading to a flood of snake oil salesmen and overhyped software tools. Much of the GEO landscape is a tangle of conflicting advice, myths, half-understood jargon and repackaged SEO advice. It can be a mess.

Even the basics are poorly understood. There’s a tendency to think of tools like ChatGPT as sponges that are constantly absorbing new information. In fact, models are trained on an enormous body of data and frozen in time until the next update—a period that can currently range from six months to 18 months for major platforms—a limitation known as the AI knowledge cutoff. For example, Claude Opus 4.8 was released in May, on data frozen in January—the model itself does not know, for example, who won the latest World Cup match.

But, savvy to the fact that users want up-to-date information, AI companies have developed a technology called retrieval-augmented generation, or RAG. It’s a fancy acronym, but all it means is that chatbots do a manual search (ChatGPT uses Bing, for example) and quickly distill the results for the user. The underlying model doesn’t absorb the information—if you ask for the latest World Cup results, it’ll look them up, but it won’t integrate them into its knowledge base and future responses.

That two-step process makes GEO wonky. Getting a company into the base training data is a longer-term project that requires building authority in trusted sources that AI companies can access. Getting into RAG results is not much different than classic SEO: It’s a chatbot that’s doing the searching instead of a human, but the same principles apply.

Measurement, too, is hard. In “classic” SEO, you could come up with a reasonable approximation of how well you were doing simply by looking at search results: If you went from being 15th to fifth on a Google search for “best interior designer St. Louis,” you knew you were doing something right. There’s no equivalent metric in GEO.

Conversations with AI chatbots are happening in private, and are highly attuned to the individual user: Two people sitting next to each other will get different results if they ask ChatGPT for the best interior designer in St. Louis—and no one outside of those two conversations will clock what was said. Those limitations make it difficult to get a real impression of how often your name is dropped by AI.

Startups like Profound have come up with complicated workarounds. Like 5W’s study, they’ll hammer chatbots with prompts (albeit at a much larger scale), and layer in data analysis on top of the results. Various tools offer up different kinds of analysis and metrics. But at the moment, it’s an inexact science.

“You can track visibility, you can track share of voice, and citations,” says Ian Lurie, a veteran digital marketing consultant. “You should do that, but [the industry] is still figuring that part out.”

There is good news. If you have already done some basic SEO work, you don’t need to drop everything and start over. “GEO is a unique thing, but I also don’t want to overemphasize the uniqueness,” says Lurie. “Ninety percent of GEO is also SEO. The best practices are nearly identical up to that last little 10 percent.”

Even better—for all the ways that GEO is genuinely different from SEO, most of the distinctions are relatively straightforward. SEO, for example, typically involves trying to optimize around keywords (like, say, “best interior designer St. Louis”—the kinds of things a user might have typed into a Google search bar). Because AI searches are conversational—putting the chat in chatbot—GEO involves optimizing around questions and answers, not mere keywords or phrases.

“People don’t go to a generative engine like Gemini or ChatGPT and type ‘e-bikes,’” explains Lurie. Instead, they’re sharing context. “They say, ‘I’m a commuter. What’s a good e-bike for me?’ With GEO, it’s particularly important [to frame your content around answering a question] because generative engines will pretty much ignore you if all you do is write to a keyword.”

Also much like traditional SEO strategy, it doesn’t hurt to tailor your copy to what people are already looking for. Jennifer Smiga, founder of digital marketing consulting firm Ultraviolet Agency, points out that often there’s a disconnect between using the “right” terminology and tapping into popular search terms. “Designers might say, ‘I want to write about boucle chairs,’ but [the general public] is saying ‘fuzzy chairs,’” she says. “I can confidently say people are not looking up ‘boucle chairs.’”

When it comes to cultivating outside links and mentions, many of the classic SEO principles apply, with a few tweaks. Reddit is often cited as especially impactful in influencing AI search—it has an agreement with Google, making it especially relevant to Gemini and Google AI overviews. Interestingly, so is YouTube. But Lurie cautions clients not to overthink where they build authority online.

“You want your brand to show up in as close proximity to all the other content on the subject as you possibly can,” he says. “The truth is, you want to post where other people go. … If you can give good advice, though, and answer meaningful questions on Reddit, absolutely do that. I would start with the places that you know people hang out.”

Despite the complexity of the underlying technology, a lot of the best practices are intuitive. Smiga’s audit process doesn’t even start with tech, but rather a long questionnaire designed to help designers and brand owners discern what truly sets them apart. Without a clear understanding of your own appeal, it’s unlikely that you’ll be able to convey it to an LLM—or the human who’s using it.

“That’s really important foundational work,” she says. “You want a singular point of view, so you’re telling all the robots the same story.”