AI Is Poised to Make Google Obsolete in Just Four Years

AI could overtake traditional search in just a few short years. 

We’re already seeing it with the incorporation of AI results in search engines, but even past that, OpenAI launched ChatGPT Search in October of 2024, which enables users to quickly get information on any topic via a chat with the large language model. And Perplexity offers a robust AI search engine that is completely up to date with current events and content.

The New Search Revolution

SEO revolutionized marketing in the early 2000s. All it took was some elbow grease, a solid SEO strategy, and patience as a site would slowly climb the keyword rankings. And of course, the world optimized its content to rank in one specific place: Google.

But now, the world is changing, and even Google itself is no longer the great source of traffic for websites that it once was. 

About 60% of all Google searches result in no clicks. That’s right. The majority of searches result in the searcher not clicking on a single link. And no link clicks means no visitors to your site.

Rand Fishkin and the good people over at Sparktoro worked with Datos to get these stats:

On top of that, they estimate that the total share of online searches done on ChatGPT is 4.33%. Compared to Google’s 83.54% market share, that doesn’t seem like a lot. But to put it in perspective, Bing’s share of search is only 1.97%. So in less than a year, ChatGPT has more than doubled Bing’s numbers. At its current growth rate, ChatGPT will catch up to Google in as little as 4 years. 

But ChatGPT isn’t the only large language model (LLM) game in town. Google’s own Gemini, Meta’s Llama, Perplexity, and Anthropic’s Claude are all growing in usage. Each of these tools could replace most of what Google’s existing search engine does.

Be Where Your Customers Are

“Marketers and PR pros face a triple threat to their playbooks in 2025,” says Daniel Lemin, Founder of One Good Brand. “First, people are leaving "big" social platforms and engagement is struggling. Second, they are shifting their attention to spaces that are private, hard-to-reach (via paid or organic presence) and difficult or impossible to monitor. And third, they're using Google less and less, turning to AI tools and summaries.”

But unlike Google, where if a user doesn’t find the answer in the first couple of links they can just keep scrolling until they find one that fits, LLMs will give a succinct and presumably believable answer to the query without much other thought by the user. 

So if you don’t show up in that initial response, you basically don’t exist.

This change in consumer behavior is going to force marketers to take up a new proficiency that many have never heard of, Large Language Model Optimization.

Showing Up in LLMs with LLMO

Organizations need to start thinking about getting mentioned in LLMs for queries that match their own offerings. The process for optimizing the chances of your organization being mentioned in an LLM is called Large Language Model Optimization or LLMO. 

LLMO will become the new SEO - or at least live beside it in the marketing stack. Currently, LLMs reference articles that rank well in search engines when they create their responses. This means that if you’re currently ranking well in search, there’s a good chance the LLM will feature you. But that may not be true in the future.

LLMs that are doing live searches are still new and are simply using an existing tool (search engines) to help pull the correct sources of information about a topic. But it’s unlikely that will stay true in the future. As LLMs evolve, SEO rankings will have less and less impact on the content of the LLM responses. 

That’s when LLMO will truly take off. And figuring out how to get content ranked in an LLM will become essential for many organizations.

How You Can Start LLMO

Right now, many of us are working on figuring this out, but as mentioned before, a big part of it does appear to be tied to search rankings. Semrush found that LLMs cited the top 10 ranking sites in Bing or Google in most responses. They also found a large correlation between higher ranked content being cited in an LLM’s response.

Another factor that will most likely impact topics being associated with organizations is mentions of your brand with those topics getting featured in large publications online. 

LLMs learn by joining different ideas when it sees training data that includes both repeatedly. It’s likely that larger publications featuring this information will carry more weight than just it being mentioned on your website.

“You have two mandates this year. A word-of-mouth strategy that compels people to talk about you in those spaces where they're turning their attention, and a refreshed content strategy that gives you a leg up on AI answers and summaries,” says Lemin.

So much of LLMO may be good old fashioned PR! 

Getting brand mentions with various topics will replace SEO best practices of anchor text and inbound linking. As LLMs ingest and train on this data, they will begin associating those topics with the brand and make it more likely that searches for those specific topics would come back with a mention of the brand.

PR may play a huge part in the next phase of marketing. Is your organization prepared?