- Kiera Carter
- Jul 29
- 4 min read
Updated: 5 days ago
Your brand's representation in large language models could mean the difference between acquiring new customers and losing them to competitors.
If navigating the new world of AI search sounds stressful, keep in mind that marketers have always fought to dominate the platforms that matter most at any given moment: magazine mentions, social feeds, Google Search, and now, LLMs.
In some ways, the fundamentals haven't changed. You want to protect your brand’s reputation, control the narrative and stand out from your competition. The million-dollar question—perhaps literally—is how?
That’s where generative engine optimization comes in. So, what should you do if an LLM is badmouthing your brand to potential customers? Or, perhaps worse, getting your brand completely wrong? Here’s a step-by-step guide to changing the conversation.
Related: Wix Studio's AI tools will transform the way you work
How to change negative brand mentions in LLMs
01. Monitor brand mentions
You can’t change the narrative if you’re not tracking the narrative, and that’s admittedly tricky in a space that’s both new and limitless in the conversations your customers could be having. (Remember the sweet simplicity of keywords?)
Consider one of these LLM visibility tracking tools. If you're a Wix or Wix Studio user, you can also tap into the new AI Visibility Overview, where you can monitor AI search performance right from your Analytics dashboard.
Still, Rebecca Tomasis, blog SEO team lead at Wix who oversees LLM optimization, says not to overlook the manual method. To start, “we identified our top branded searches, and we manually checked them on ChatGPT and Perplexity, two of the biggest LLMs right now.”
They found some issues with a brand they had recently launched. “When the LLMs couldn’t find anything about the new brand, they’d refer to Wix for information, and they couldn’t always distinguish between the different brands,” she says.
So, then what?
02. Give the LLM feedback
Crystal Carter, Head of SEO Communications at Wix, likes to give a thumbs down to incorrect information. “LLMs are trained on human feedback,” she says. “If you see them say something factually inaccurate, you can give the statement a thumbs down and correct it.”
The conversational nature of LLMs also presents a unique opportunity to dig deeper. “If you don't think you're being ranked fairly, you can just ask for the reasoning,” Zack Notes, founder at Sandbox SEO, told us in Is LLM optimization the same as SEO? “Sometimes they will even cite a specific review or article that tipped their opinion. This breaks open the black box of traditional search.”
03. Update your “about” page and other high-impact content
Tomasis and her team tackled the issue on two fronts:
The brand’s owned content (blogs, landing pages, etc.)
External content (affiliate content, media mentions, etc.)
“It’s easier to change the conversation when you’re showing up incorrectly—compared to not showing up at all—because that means the LLM already knows your brand,” Tomasis says. “The information is built on SEO. It’s crawling blogs. It’s processing marketing campaigns. The LLM knows your brand, and it’s pulling the information.” (If you’re not showing up at all, then you have an SEO problem. We’ll cover that in a future post.)
That means you can write blog posts and update your About page to reflect the brand story you want to tell. “We see our updated blog posts rank in Google within 24 hours, Perplexity 24 hours after that, then ChatGPT two or three days after that,” Tomasis says. “In general, once we get cited in the LLMs, then the conversation changes almost immediately.”
04. Reach out to external partners
But—and this is a big but—you need to look at what everyone else is saying about your brand online, too. “LLMs are such a holistic thing, so the messaging doesn’t necessarily stick if all the other citations still have outdated or incorrect information,” Tomasis says. “Sometimes you’ll check an LLM, and it’s pulling more strongly on an affiliate that’s outdated. Everything across all your web assets needs to be updated and aligned.”
Unfortunately, this means another manual task for your team. Reach out to affiliates and media partners to ask them to update the incorrect content. Keep in mind: These people are swamped and they don’t work for you, so make their jobs as easy as possible with a super-clear request and respectful suggestions.
05. Approach it as a brand problem, not an LLM problem
Most LLMs have a way to report or flag misinformation. Should you go this route to combat misinformation about your brand? Well, that’s not really in the spirit of the flag, which is meant to protect users from dangerous hallucinations and advice that could have harmful effects, like wrong health information.
Tomasis has a fair way of looking at this: “If LLMs are saying something incorrect or unflattering about your brand, then it’s a brand problem, not an LLM problem, and it’s the marketer’s mess to clean up,” she says. “As long as old and incorrect content still exists online, and as long as new content is missing, it won’t tell the story of your brand.”
