Ever since the first iteration of ChatGPT arrived, marketers have been obsessed with a couple of key questions: How can I get my brand(s) to show up in these tools, and how long do I have until they take my job?
Three years on, we’re finally starting to get some answers—and the emerging consensus is a little different than some of the initial hype may have suggested.
GEO Lessons From a Failed Startup
This week, we got some evidence that there’s no “one weird trick” for improving your GEO performance.
And the call came from inside the house: Benjamin Houy, the founder of Lorelight, a “proactive AI brand monitoring tool,” published a blog post announcing that he was shutting the startup down. His reason? Here are Houy’s own words (emphasis added):
After analyzing hundreds of queries across ChatGPT, Claude, and Perplexity, I found that brands with high AI search visibility all had the same characteristics:
- Quality content that genuinely helped people
- Mentions in authoritative publications
- Strong reputation in their space
- Genuine expertise and thought leadership
Sound familiar?
It should. Because it’s the exact same stuff that’s always worked for SEO, PR, and brand building.
There was no secret formula. No hidden hack. No special optimization technique that applied only to AI.
The AI models are trained on the same content that builds your brand everywhere else. They cite the same authoritative sources. They reference the same trusted publications.
There’s no such thing as ‘GEO strategy’ or ‘AI optimization’ separate from brand building. At least not for the vast majority of brands.
Real World Examples
I will pause here and note that there may well be some confirmation bias on my end—Houy’s conclusions happen to agree with my own understanding of the current landscape. But that’s an understanding I’ve arrived at independently, in part by looking at client results.
For example, I recently ran a series of AI searches for a client in the mortgage industry related to content we produced for them in Q2 (subject: mortgage considerations for people going through divorce). While we optimized the content for SEO, we didn’t do anything special to help it show up in AI. And yet, just a couple of months after posting, it was not only ranking for traditional organic search but was being cited by multiple AI tools alongside content from some of the best-known firms in the industry.
A second example comes from a different client, in the commercial glass industry, who recently informed us that they’d spent an hour trying to uncover content areas related to their products where AI tools didn’t cite their content. They failed, and noted that some of the blog posts cited by AI search engines were published several years ago. While that means they couldn’t have been “optimized” for AI, they are among the client’s best-performing posts in organic search.
The Human Advantage
If there’s one thing I’ve learned in more than a decade of grappling with SEO, it’s that there are always new tricks, but they don’t work for long. Every “can’t-miss” tactic I’ve seen eventually fades, either because it’s overused, gets penalized in the next update, or the algorithm itself moves on.
The reason is simple: algorithms—whether they power Google or generative AI tools—are always evolving. What works today might be irrelevant by the next quarter, and this pace of change is only accelerating as AI becomes more central to how people discover information.
That’s why the fundamentals matter most: creating original, valuable, well-researched content that answers real questions and builds trust. But fundamentals alone aren’t enough—you also have to keep an eye on where things are heading. As AI search evolves, so will the strategies for optimizing visibility. Our approach is to double down on what’s proven, while continually scanning the horizon for the next shift.
In this landscape, there may never be a permanent shortcut. But staying focused on the basics—and staying nimble as the rules change—gives your brand the best shot at being found, in search engines and AI conversations alike. |
|
