A little over a year ago, I wrote an article titled “The future of SEO in an AI-powered world.”
I wanted my opening line to be dramatic, so I wrote “The Google interface we know today will be gone in less than three years.”
At the time, this line sounded outlandish. Could the world’s most popular website, with close to 2.5 billion daily users, just vanish?
It’s a year later, and I find myself overwhelmingly turning to LLMs over Google for all kinds of information.
In my case, the prediction of three years was an understatement. I suspect the same is true of many of you reading this.
Is Google dead?
Not quite. To paraphrase Mark Twain, news of Google’s death has been greatly exaggerated.
As industry leaders like Rand Fishkin have pointed out, Google is more popular today than it ever was.
And he’s right that even as the runaway market share leader in AI, ChatGPT‘s usage is still a drop in the bucket compared to Google’s.That said, it’s important to remember history.
Henry Ford introduced the Model T in 1908. `
The horse population of the United States reached its all-time peak of 25 million horses in 1915.
It wasn’t until 1920 that horse sales began to decline.
Here’s a more recent and relevant example.
Google launched in 1998. Yahoo reached its peak of users in 2007, but its usage began to collapse in 2008.The Google-Yahoo dynamic is an interesting one because it parallels today’s situation.
Yahoo was a web directory. Google was a search engine.
While Google battled with other search engines like AltaVista, Execute, and Lycos for supremacy, Yahoo was the dominant place to go to find websites.Google went on to overtake Yahoo in 2006 and never looked back.Today, Google is a search engine.
ChatGPT is an “answer engine.”
ChatGPT is battling with other answer engines like Perplexity, Claude, and Grok for supremacy, while Google is the dominant place to go for information.
Dig deeper: Answer engine optimization: 6 AI models you should optimize for
Search engines vs. answer engines
For years, the phrase “search engine” meant “someone searching for a website,” and Google has long been dominant in that space.
As far back as 1996, Ask Jeeves toyed with the concept of an “answer engine,” where you could ask a question and get a natural language response.
But the technology wasn’t there yet. Starting around 2012, Google started toying with SERP features like answer boxes, knowledge panels, and featured snippets, which presented “answers” on the Google SERP without users needing to click through to a website.
As the years passed, they added more “answers” to the SERPs from local results to flight information to music lyrics.By 2019, this rise in “zero-click searches” became a topic of debate.
Some search experts viewed it as a money grab by Google, while Google itself framed it as continuing to evolve its site to serve its users.
It turns out both were right.What was really happening?
Google was correct that for certain kinds of queries, users don’t want to click through 10 results.
They just want the answer.
As the default search engine on 90% of the world’s browsers, they had the luxury of being able to slow-walk the transition from search engine to answer engine.
But in November 2022, OpenAI forced their hand.
I remember in 1999 when people talked about Google, they said things like “it’s creepy,” “it’s like it’s reading your mind.”
For the first time in over 20 years, I heard people saying the exact same things about ChatGPT.
ChatGPT was the realization of what Ask Jeeves had aspired to do decades earlier.
When I wrote about the “butler” in my article a year ago, it was Jeeves I had in mind.
Are search engines dead?
I hear this question a lot, and it comes from a fundamental misunderstanding of what “search” means.
“Search” simply means that human beings are looking for something.
A “search engine” is an application that helps them find it.
The card catalog in your local library is a “search engine.”
Most people use:
YouTube as their “search engine” to find videos.
Amazon as their “search engine” to find books.
Almost every brand has its own “search engine” on its site, called internal search.
Google is the primary engine for searching for websites.
AI is not going to replace any of those things.
Ironically, AI will become one of the top consumers of search engines as it – joining us humans – searches for knowledge.As for Google, there are certain kinds of “searches” that it’ll continue to dominate.
The types of searches that Google will dominate
Navigational searches
Of the top 25 searches in 2024, 18 (72%) involved people literally typing in a brand name to find a website.
Let’s face it, we humans are pretty lazy, so if we can save four keystrokes by not having to type “dot com,” we’ll do it.
Real-time or fresh information
People have grown accustomed to turning to Google for fresh information, such as sports scores and weather.
LLMs often have a cutoff date or delayed access to real-time data.
Local information
Google has done an impressive job of dominating local search.
Thanks to its double-dipping approach, it serves:
Its own first-party content first.
Third-party content like Yelp and TripAdvisor underneath.
Official and authoritative sources
People often use Google to find specific documents or tools that are buried or difficult to find directly on a website.
Examples include:
IRS forms.
College applications.
Public school calendars.
Customer service phone numbers.
Tools such as USPS and UPS tracking.
In a lot of cases, these sites don’t do a great job at UX and internal search, so Google (and the site: operator) becomes any site’s de facto on-site search tool.
Visual search
For image searches and reserve image searches, it’s hard to see anyone challenging Google.
Ironically, the more pervasive AI-generated imagery becomes, the more people will flock to Google Images to look for authentic human imagery. (If Google can figure out how to allow users to differentiate between the two.)
If you look at the top 100 searches on Google in 2024, you’ll see that the vast majority of the keywords on that list are in one of these five categories.
They add up to nearly 14 billion searches.Will AI challenge Google on any of these? Probably not. Google’s greatest threat right now is the government.
If the DOJ or the EU decides, for example, that Google must divest the Chrome browser or give Yelp equal visibility as local packs, that could jeopardize some of these top searches.
Get the newsletter search marketers rely on.
See terms.
7 ‘searches’ that LLMs will dominate
Do you see what all the searches above have in common?
They’re all searches for “stuff.” Specifically, “stuff” that exists on other sites.This was in keeping with Google’s original mission statement:
To organize the world’s information and make it universally accessible and useful.When you look at the mission statements of AI companies (including Google’s own DeepMind), the mission statements are a bit different.
“To ensure that artificial general intelligence benefits all of humanity.”
“To build AI responsibly to benefit humanity”
“To cater to the world’s curiosity”
“To ensure AI systems are safe, beneficial, honest, and harmless.”
“To advance our collective understanding of the universe.”
While Google’s original mission was about organizing and serving information, AI companies are about providing knowledge.
But if you carefully observe the difference between queries on Google and prompts on chatbots like ChatGPT, Claude, Perplexity, Grok, and Google’s own Gemini, you’ll see stark differences as people get used to the new media of AI.As people become accustomed to AI, prompts will become less like short and choppy phrases that resemble Google searches and more conversational.While it’s difficult to predict the future, here are the kinds of “searches” where it looks like LLMs will shine over the traditional Google interface.
1. Research
I casually mentioned above what the horse population was in 1915.
No, I’m not Ken Jennings. I just asked ChatGPT and it gave me the answer in under five seconds.As more people learn how to write “prompts” instead of “queries,” more people will realize how easy it is to sound smart without really being that smart.Before AI Overviews, a Google search like “horse population after cars” would have sent me wading through Reddit threads, Quora posts, and a dozen blogs before coming close to finding an accurate answer.
Dig deeper: Want to beat AI Overviews? Produce unmistakably human content
2. Planning
I’ve been using ChatGPT for a while now, but this was the “killer app” that turned me into a daily active user.My daughter is in the fourth grade, which means she gets a free Every Kid Outdoors pass to visit any National Park.
My daughter’s teacher told us how much she loved Zion National Park, so we decided to go there.Our daughter’s spring break was a week, so on a lark, I asked ChatGPT how many National Parks I could visit in a week.
It not only educated me on Utah’s “Mighty Five,” it also gave me an itinerary with driving times and key sights to see.
After a little more conversation, I had my whole trip planned out. And it turned out to be one of the best family vacations we’ve ever had.Now, imagine this for sprint planning. Or event planning. Or planning a wedding.
As long as you give it the right inputs, you’ll be shocked at how much time it can save you.
3. Translation
If there’s one “killer app” that should have Google shaking in its boots, it’s translation.
“Translate” and “Google Translate” were the fourth and 10th highest searched phrases worldwide in 2024.If you ask anyone who does a lot of translation, chances are they will tell you that while Google Translate is okay, ChatGPT is revolutionary.Here, I’ve asked Google Translate and ChatGPT to translate an early draft of this article into Traditional Chinese.
If you show these to someone who reads Mandarin, they’ll be able to tell you right away that:
The first is more accurate and idiomatic, and reads much more naturally overall.
The second contains multiple errors and sounds more awkward.
But don’t take my word for it. Try it for yourself.
Ask Google Translate and ChatGPT to translate something into a language that you or someone you know can read, and see how well they do.
ChatGPT’s Chinese translation
Google’s Chinese translation
My wife, who’s a native speaker, agreed that ChatGPT was better.
Just for fun, I asked Grok to corroborate:
4. How-tos and step-by-step instructions
This is another area where AI has the potential to outperform traditional Google search.
I recently bought a Brother color laser printer, and as a lifelong HP user, I was confused about how to change the toner.
Googling “Brother MFC3780CDW toner” led me down a rabbit hole of shopping results and Brother’s own confusing site.
The key difference is that Google surfaces various sources written for broad audiences, while AI delivers answers tailored to your specific needs.
Google’s decision to deprecate HowTo schema and remove “how-to” rich results in 2023 was likely an admission that users didn’t want that kind of content cluttering the SERPs.
That said, AI still can’t fully replicate image or video instructions (yet), so there’s still plenty of room for creators of “how-to” content to thrive.
5. Editing
One of AI’s most powerful applications is writing and editing.
AI can handle a range of editorial tasks:
Fixing grammar and spelling.
Drafting outlines.
Rewriting paragraphs.
Even producing entire pieces.
I’ve used it myself to tighten long-winded passages, consolidate scattered ideas, and strike a better balance between detail and conciseness. This very passage was refined with the help of ChatGPT.But while AI is useful, I’m already seeing it misused – often lazily and foolishly.
All across the web, you’ll find writing that looks polished: it’s grammatically perfect, impressively worded, and feels substantial.
But if you read closely, they seem soulless.
Google has told us that the secret to good writing is E-E-A-T (experience, expertise, authoritativeness, and trustworthiness).
But how do you differentiate yourself in a world where AI has made E-E-A-T a commodity?
You double down on what only humans can do.
Be Original: AI can’t truly innovate – it can only remix what’s already out there. New ideas still come from you.
Be Practical: If your content doesn’t help people, they won’t engage. Write to solve real problems.
Be Relevant: AI can’t share personal stories or lived experience. You can, and that’ll help your readers relate to you in ways AI writing cannot.
Be Authentic: AI-generated content often sounds generic. Let your voice and perspective shine through.
Be Human: In a world of flawless machine-written text, it’s the human quirks – your tone, humor, even imperfections – that will be the key differentiation.
Yes, that spells O-P-R-A-H, but I promise that’s just a coincidence.I know some writers who refuse to have anything to do with AI, and I know others who have AI do all their “writing” for them.I fall somewhere in between.
I treat it like a creative partner – I’ll use it to brainstorm, generate outlines, help me think of words, or reshape clunky drafts.
But no matter how AI gives me, I’ll go through as many “final passes” as I need until I know it’s my voice and my “soul” coming through, and not a robot’s.
I’ll make sure to include insights, stories, and observations beyond what AI can provide.
6. Generative images
AI-generated images have come a long way from the days of seven-fingered models and the uncanny valley.
Today, DALL-E is the most accessible option, Adobe Firefly is popular among Creative Cloud users, and Midjourney is arguably the most photorealistic.
But as AI imagery becomes more widespread, you may start to notice the same problem I mentioned with AI-generated copy: much of it feels soulless.
We’ve seen this before.
When free stock photo sites exploded in popularity a few years ago, the same images began appearing everywhere.
It quickly became obvious which brands were cutting corners and which ones were investing in visuals that truly connected with their audience.
The same pattern is unfolding with AI.
Some will settle for generic results from a quick prompt, just to fill space.
Others will take the time to refine their prompts to get something truly compelling.
And the more sucessful of all will be human artists who follow O-P-R-A-H (You get a click! And you get a click!)
7. Distilling large amounts of data
Something else that AI can do uncannily well is take large swaths of information and distill it into a clear, concise summary that’s tailored just for you.
This can be useful for providing summaries of things like:
Government documents.
Terms of service.
Legal agreements.
Long emails.
Meeting transcripts.
Where it gets interesting is once you’ve trained an AI on a large document by uploading it, you can start asking your AI questions about it.
What does this all mean for me as a marketer?
You might be thinking, what relevance does any of this have to me as a search marketer?
The answer is – everything.
When I researched horses in 1915, I became so immersed in the topic that I asked ChatGPT where it got its information, then dove into its sources.
When I planned my vacation, I asked ChatGPT about specifics, from hotels to restaurants to dark sky experiences. It recommended businesses, and they got my money.
When I looked up how to change my toner, I asked ChatGPT to give me suggestions for generic brands to save money. It gave me a list.
The difference is this: Unlike search engines, where you fight for rankings, with AI, you’re fighting for a spot in the AI’s consciousness.Here’s how to do that.
Don’t take your eyes off Google
As shiny as AI feels right now, let’s not forget Fishkin’s point:
Google is still the primary way people search – for both information and websites.
Ignore it at your own risk.
I think Google will continue to grow. But AI will grow faster.In Google’s early years it relied on DMOZ to help power a lot of its results.
In the same way, AI will rely on Google and other search engines for its “research.”
Make your brand indispensable
This is the most important move you can make for the AI era.
People don’t search for “best phone” to get to Apple, or “online bookstore” to find Amazon.
They search for the brand.
Too many companies invest in search and ads without first ensuring that their products, services, and support actually meet customers’ needs.
SEO doesn’t rescue a weak brand.
A strong brand makes SEO (and AI) work better.
Instead of building links and stuffing keywords, build trust.
Be the brand that humans think of first. Because AI won’t “think” of you unless people already do.
Dig deeper: The new SEO imperative – Building your brand
Make sure AI can see you
People often ask me what tools they can use to track AI visibility.
Good news, you already have them:
ChatGPT.
Perplexity.
Claude.
Gemini.
Grok.
Start typing your brand questions into these tools. See what answers show up.
If you’re being cited, great.
If not – or worse, if the AI starts hallucinating (it happens) – you’ve got work to do.
Check that your content is crawlable, well-organized, and actually answers real user questions.
Don’t block AI bots. And make sure your own on-site search works as well as (or better than) Google for your brand.
Dig deeper: 3 ways to optimize for AI search bots
E-E-A-T still matters
I was never a big fan of the acronym E-E-A-T because a lot of SEOs tried to game it rather than really understand it.
The problem is that, with search engines, only one can be the top “expert” and “authority” for any given head term.
But as people get more used to giving “prompts” instead of “queries,” long-tail demand will explode.
That’ll open up new and limitless opportunities for you to be the expert on something specific.
Double down on your humanity
You might notice something throughout this article.
I included personal stories, real examples, and even bad dad jokes. Why?Because as AI-generated content floods the web, what will stand out is content with individuality and a “soul.”
Customers will crave it. AI will learn from it. If you keep delivering it, both will trust your brand more.
Dig deeper: The art of AI-enhanced content: 8 ways to keep human creativity front and center