Takeaways from Ethan Smith's talk at Ahrefs Evolve
Zach and I went to the Ahrefs Evolve conference in San Diego. One of the most interesting presentations was given by Ethan Smith of Graphite, a research-driven SEO agency.
He talked about brand search in AI.
When someone asks an AI assistant, "What do you think about [your brand]?", the answer they get may be the most important piece of marketing you never wrote.
His session dug into how search is changing now that people get answers from AI instead of a page of blue links. While he talked, we scribbled four lines on the whiteboard, and they stuck with us enough that we rewatched the session recently. Those four lines are the headers below. They're simple, but they reframe how every company should think about its reputation.
AI tells your brand story
For two decades, brands controlled their own narrative online. You wrote the About page, the product descriptions, the case studies. If you ranked, your words were what people read.
That's shifting. More buyers now skip the search results and ask an AI tool directly. When they do, the AI doesn't hand them your homepage. It tells them a story about you, in its own words, assembled from whatever it has learned.
"Tell me what you think about my brand"
Try this yourself. Open ChatGPT, Gemini, or Perplexity and type: "Tell me what you think about [your brand]." Then ask about your competitors.
What comes back is, in effect, your brand story as the market hears it. It might be accurate. It might be outdated. It might lean on one bad review from four years ago or a forum thread you've never seen. Either way, that answer is shaping opinions before a prospect ever lands on your site.
AI summarizes the wisdom of the crowd
Here's the part that matters most. These models aren't forming opinions from scratch. They summarize what people have already said: reviews, Reddit threads, YouTube videos and comments, news coverage, comparison posts, and yes, your own content.
Your own profiles anchor a lot of that opinion. The About pages and descriptions on your YouTube channel, Instagram, subreddit, Tumblr, Facebook page, and blog all feed the picture, and those are the pieces you fully control. Make sure every one of them says clearly who you are and what you do.
AI is a mirror held up to the crowd. It reflects the collective conversation about your brand, condensed into a few confident paragraphs. And because your own content is part of that crowd, what you publish helps shape the answer.
The more people say something, the more it will say that thing
This is the line that ties it all together. Repetition is a signal. If a hundred people across a dozen platforms say your product is reliable, the AI will say your product is reliable. If the dominant thread is that your customer service is slow, that's what it will repeat, to everyone who asks.
That makes brand building a volume and consistency game across the whole web, not just on your own domain. A few practical implications:
- Audit your AI reputation. Regularly ask the major AI tools about your brand, your category, and your competitors. Track how the answers change over time.
- Show up where the conversation happens. Smith pointed to Reddit, YouTube, and backlinks as key levers in the new search era. If people discuss your category there and you're absent, someone else is defining you.
- Earn mentions, not just links. Getting your brand named in credible articles, reviews, and roundups builds the body of evidence AI draws from.
- Say the same true thing, often. Decide what you want to be known for and reinforce it consistently in your content, your PR, and your customer experience.
- Fix the real problems. If the crowd is complaining about something legitimate, no amount of content will drown it out. Fix it, then help happy customers share their experience.
You are training the models
Every piece of content about your brand becomes part of what AI learns from. Your blog posts, your product pages, and every article that mentions you are material the models read and summarize. In that sense, you are helping train them, whether you plan to or not.
That makes two things important. First, publish content on your own site that says clearly what your brand does and what it stands for. Second, get that same message published across the web, on sites the models already read and trust, so your own claims are backed up by other voices.
This is a big part of what we do at Page One Power. Our Media Blitz turns a data study on your site into coverage syndicated across hundreds of news outlets, each with a link back to you. Our PR Drop follows with brand mentions across 500 to 800 outlets. Think of it this way: the Media Blitz builds links to your site, and the PR Drop spreads your name. Together they put your story in front of AI in many places at once, and keep it there.
Consistency matters as much as volume. One campaign gives you a short burst of attention. Steady coverage and a regular flow of links back to your site keep validating what you say about yourself, and that repeated validation is what the crowd, and the AI summarizing it, picks up on.
The bottom line
Traditional SEO was about getting found. AI search is about how you're described everywhere. Your brand story is no longer something you publish once. It's something the internet tells collectively, and AI repeats.
The good news is that it isn't out of your hands. Every review you earn, every helpful video, every honest mention on a forum adds to the chorus. Give the crowd something good and consistent to say, and the machines will say it too.
