Offer: 2 weeks instead of 1 free with Semrush

Over ¼ of Americans say they’re ditching Google searches in favor of AI.

So that’s it, wrap up your SEO strategy, it’s over. 

Just kidding. If you unpack the stats a little (which we did), you’ll find that the shift, while real, is nowhere near what the hype would have it seem. 

Thanks to insights by Semrush and a few other sources, we now know more about who is searching with AI and why, and how you can capitalize. 

Get up to speed below.

“Classic” for all the wrong reasons.

Coca-Cola recently rolled out an AI-driven ad campaign, celebrating mentions of the drink in works by famous authors. However, the AI glitched, and the brand included a book that doesn’t exist at all.

Which famous author did they attribute the book to?

• JK Rowling

• Stephen King

• J.G. Ballard

• Roald Da

Scroll to the bottom to find out!

27%.

The number of Americans ditching Google search in favor of AI search tools.
But don’t kill your SEO strategy just yet.

A 2024 survey from Future (publisher of TechRadar) says more than 1 in 4 Americans are making the switch from traditional search to AI tools like ChatGPT.

Does that mean 1 in 4 searches are now made through AI? Not quite. 

• If you look at search traffic as a whole, AI chatbots pull just 3% of total searches (OneLittleWeb study based on Semrush data)

• Only 10% of U.S. internet users start their searches with GenAI (Semrush x Statista)

• While 71.5% of users say they’ve tried AI for search… only 14% do it daily (Search Engine Land)

In other words: People are curious—but not committed.

Still, the shift is real. Semrush estimates 90 million Americans will use GenAI as their primary search tool by 2027, up from 13 million in 2023. Semrush internal studies estimate this will be the tipping point where 75% of Search-driven revenue will come from LLMs.

But what makes an LLM choose to mention you over another brand? SEO-type rules still apply: content quality, site architecture and the like. Read on to learn more 👇️ 

💡 Check your website’s SEO here.

The data says it’s time to optimize for AI

Search engines, whether they’re the traditional keyword-based kind or the sorta-thinking-LLM variety, absolutely do not share the secret of their ranking algorithms.

They don’t want you to publish optimized content. They want you to publish “good” content that increases the value of their search tool. 

But there are 1.1 billion websites on the internet, and more than 177K added to that number every day. Crawlers and algorithms can’t actually find the best content in that heap — so they tend to show the most optimized, like it or not.

So, how do you optimize for AI search if ranking is such a closely guarded secret?

Semrush analyzed hundreds of thousands of searches to get that answer for you.

📑This issue shares the highlights. Want the detailed report? Get it here.

First look: what the data shows

Google currently dominates search, whether it’s AI-answered or not. Therefore, the best place to start optimizing for AI is in Google’s AI overviews.

Here’s what we know about them: only 13.14% of searches show an AI summary. 

But that number is trending higher. Since January 2025, the percentage of searches that generate an AI summary has more than doubled.

These searches tend to be zero-click topics, which means people get the answer they want from the SERP without ever visiting your site. 

Zero-click doesn’t mean you don’t want to show up here. It just means that you should think of AI optimization in terms of visibility. Whenever you’re not mentioned, your competitors are.

The numbers change rapidly, but early trends point to specific types of content being more AI-friendly.

Optimize for AI if you publish any of these types of content

The same things show up consistently in AI overviews:

• 88.1% of queries that trigger an AI overview are informational content.

• AI overviews are most common on low-volume keywords with under 1,000 monthly searches.

• 95% of the time, there are no PPC ads for keywords that trigger an AI overview.

Semrush also found that certain topics have a much higher likelihood of AI search features.

• Health

• People and Society

• Science

• Food and Drink (on mobile only)

If you’re publishing informational content around one of those topics, it’s wise to optimize for AI.

How?

Cracking the code: how to get included in AI

SEO agency owner Jake Ward breaks all the pieces of AI optimization down into distinct parts:

#1: SEO (search engine optimization) - what gets you ranked in top organic spots?

Keywords, content, and backlinks. You know the drill.

#2: AIO (artificial intelligence optimization) - what helps AI models learn about your business?

In Jake’s expert opinion, your content must become part of the AI training data. 

Structure your web data in ways that are easy to read by machines (meta tags, schema, etc). Then, get your content into places where LLMs learn like GitHub, Wikipedia, and Common Crawl.

#3: GEO (generative engine optimization) - what gets your name mentioned in generated text?

Well-structured, factual, clearly formatted content is a lot easier for an AI to serve.

#4: AEO (answer engine optimization) - what wins AI snippets and answer boxes?

FAQ schema and question-answer structured content help a lot.

Caveat: none of this is set in stone.

AI search is still in its experimental phase. It will evolve, and that could include drastic, sweeping changes to the way things work.

Make sure the optimization you do now doesn’t get thrown out if (when) that happens.

Every time there’s a major algorithm update, the goal is the same: clear out all the stuff that’s highly optimized, but doesn’t actually serve the end user. If you always optimize for machines with the human reader in mind, you’ll be okay. 

Mostly, anyway.

A/S/L? Age, sex, and location influence your AI interactions

And demographics skew responses. So be careful out there.

Different generations use ChatGPT in different ways, says OpenAI CEO Sam Altman.

Although he caveats this by calling it a “gross oversimplification,” the age-related trends are fascinating:

• Older folk use it to replace search tools

• 20-30-year-olds use it as an advisor

• People in college use it as an operating system (complete with complex prompts and file uploads)

• College kids also use it to make life decisions from health to relationship tips — although experts are divided on whether that’s a good idea. One MIT study calls LLMs “inherently sociopathic.”

But usage isn’t all generational. Other factors come into play:

• Men make up 70% of desktop users (via Semrush and Statista) and are the majority in AI apps (Appfigures)

• Almost half of ChatGPT users have completed compulsory education only

• Non-native English speakers are more likely than native speakers to use ChatGPT frequently for writing (via ScienceDirect)

• The US represents over half of global spending on AI apps (via Appfigures)

So what?

First, you’ve got some basic audience insights into AI Search users. So that’s handy.

Second… you know AI’s users impact its responses. 

If most users are men, for example, the platforms become optimized for male-centric behaviors, preferences, use cases, and more. 

Be on the lookout for biases. ChatGPT doesn’t cancel out the need to think critically. It enhances it.

Further Insights from Semrush

Their analysis of 80 million clickstream records, combined with demographic data and traffic patterns, reveals three key changes in online content discovery:

• Traffic Distribution: ChatGPT drives notable traffic to educational resources, academic publishers, and technical documentation, particularly compared to Bing.

• Query Behavior: While 30% of queries match traditional search patterns, 70% are unique to ChatGPT. Without search enabled, users write longer, more detailed prompts (averaging 23 words versus 4.2 with search).

• User Base: ChatGPT shows higher representation among students and younger users compared to Google's broader demographic distribution.

👉️ Read the full brief here.

Written by Amy Hawthorne and edited by Catherine Solbrig.