AI & SEO: How to Prepare in 2025

When it comes to SEO and AI in 2025…

AI is a short-term opportunity. But a long-term threat.

On one hand, AI-powered tools let you streamline SEO tasks, automate grunt work, and scale like never before.

On the other hand? AI might be the biggest existential threat SEO has ever faced.

Let’s break it all down.

AI Is an SEO Cheat Code (If You Know How to Use It)

For years, SEO required a ton of manual work.

Keyword research, clustering, content writing, on page optimization, link building—it was a grind.

But with AI?

You can do these tasks faster, cheaper, and (in many cases) better.

For example: Keyword Research

Old way: You’d fire up a keyword tool, plug in some seed keywords, export a list, and then manually brainstorm related topics.

New way: You just ask AI.

Ask ChatGPT or Claude, “What are some topics my audience cares about?” and you’ll get a list of relevant keyword ideas in seconds.

No brainstorming.

No spreadsheets.

Just results.

To be clear: there’s still a place for using a keyword tool for uncovering what people search for.

After all, a tool like ChatGPT isn’t going to give you key stats around each keyword (like search volume and keyword difficulty).

It’s more that brainstorming keywords (which can be a tricky and time-consuming first step) is streamlined like crazy thanks to AI.

How about another example? AI-Powered Topic Clustering.

In the past, clustering keywords into topics required a mix of tools and manual analysis. You had to:

• Find keywords you already ranked for
• Analyze search intent
• Group related topics together

Now? You just upload a CSV and let AI do the clustering for you.

There are even a few custom GPTs dedicated to this task:

And this is just the tip of the iceberg.

AI can also help with content briefs, meta descriptions, internal linking strategies—even outlining entire blog posts (if done correctly).

In short: it’s never been easier to rank more with less effort.

Sounds amazing, right?

Well… here’s the HUGE catch.

LLMs are Eating Search

SEO has always been about playing the Google game.

You figure out what Google wants. Optimize for it.

And generate lots of traffic, leads and customers.

But what happens when people stop using Google?

Because let’s be real—it’s already happening.

People are increasingly skipping Google entirely and going straight to LLMs like ChatGPT, Claude, and Perplexity.

In fact, a survey of over 1,300 Americans by Evercore analysts revealed that 8% of respondents now use ChatGPT as their primary search engine.

(This was up from only 1% in June 2024)

And another industry study found that Gen AI search engine use is up 5.25x since January 2024.

Right now, this shift is mostly happening among marketers, founders, and people in tech.

But give it 5 to 10 years, and it’ll be mainstream.

That’s a HUGE problem for SEO.

If fewer people search on Google, then organic traffic shrinks.

Which means rankings become less valuable.

And all the AI-powered SEO tricks in the world won’t matter if there’s no one searching for your content.

But even if Google does survive, AI is still a massive threat.

AI Overviews Are Featured Snippets on Steroids

AI Overviews (formerly SGE) are already changing the SERPs. Instead of clicking on a site, users get AI-generated answers directly on the SERPs.

Which means? Fewer clicks for websites.

One early study found that organic CTR dropped by nearly 4x when there was an AI Overview in the SERPs.

On the bright side, organic CTRs actually increased when the site was listed as a source in the AI Overview.

Now, to be fair, this shift hasn’t fully played out yet. From the analytics I’ve seen, traffic is still stable (or even up in some cases) for many publishers.

But if Google starts prioritizing AI answers over organic results, the old #1 ranking could get pushed down to the new #5 or #6—buried under ads, AI snippets, and other SERP features.

And when that happens, even the best SEO strategy won’t save you.

Now it's time for me to show you how to prepare for an AI-focused SEO future.

Optimize for Google, But Prepare for the Future

As I said earlier:

AI is a short-term opportunity. But a long-term threat.

Which means Google isn’t dead yet.

In fact, traditional SEO still works incredibly well.

For example, Backlinko’s organic traffic continues to grow:

That said, you absolutely need to prepare for the AI-driven future.

But that doesn't mean you need to abandon Google altogether.

Instead, the smart move is to milk SEO for everything it’s worth while shifting resources toward future-proofing your strategy.

SEO Still Works—Here’s the Proof

Despite all the noise around AI, Google search is still an insanely powerful, high-ROI platform.

Case in point? Exploding Topics.

We target a tech-savvy, early-adopter audience.

These are the exact type of people you’d expect to be ditching Google for AI tools like ChatGPT and Perplexity.

And yet?

Our organic traffic has seen steady growth over the last 2 years. 

And we’re not alone.

Across the board, many sites (especially trusted, authoritative brands) are still thriving in organic search.

So if your instinct is to pivot away from SEO completely, don’t.

Otherwise, it’s like you’re walking away from a gold mine because you heard it might dry up 5 years from now.

Now let’s get into some tactical things you can do to optimize for Google and LLMs.

1. Continue to Build Brand and Link Authority

If you want to rank in Google, you need backlinks, brand searches, and trust signals.

In fact, Semrush’s most recent Google Ranking Factors study found a strong correlation between Authority Score and rankings.

That’s been true for years. And it’s not changing anytime soon.

But what is changing is AI models determine authority.

Why Authority Matters More Than Ever

Think about it. Both Google and AI-driven search engines like Perplexity face the same challenge:

How do you separate reliable, high-quality sources from low-value, untrustworthy ones?

To do that, they need signals to indicate which sources should be trusted, cited, or ignored.

Google has used backlinks, brand mentions, and engagement signals for years to determine authority.

And AI models will likely use the same signals that Google uses.

After all, LLMs are trained on an enormous dataset.

So they need to decide which sources to prioritize when answering a question.

The best way to do that? Authority signals.

How to Build Authority For Google and LLMs

The best way to approach SEO today is to double down on authority-building strategies that work for search engines and LLMs.

1. Earn high-quality backlinks. Traditional link building still matters. Google still uses it, and LLMs will likely consider links as a trust signal when determining which sources to cite.
   
   For example, when I ran Backlinko we would publish a ton of stats-focused content (aka “Reverse Outreach”) to build links.

   

   And the Semrush team that now runs the site has continued to use this strategy to 500+ backlinks each month.

   

2. Build a recognizable brand. Brand searches are a direct signal of authority. And they’re hard to fake. If people are actively searching for your company by name, that’s a strong indicator that your content is credible.
   
   With Exploding Topics, building a strong B2B brand has been part of our approach from Day 1.
   
   Which is one of the reasons that we send out a newsletter each week that most other companies would charge for.

   

3. Create original research and insights. AI-generated slop is everywhere. This means original data, expert opinions, and unique insights are more valuable than ever.
   
   The content that we publish at Exploding Topics is light years beyond what an LLM could produce.
   
   Instead of summarizing what’s out there, our stuff brings genuine insights to the table:

   

4. Get cited by reputable sources. Just like backlinks, mentions in authoritative publications help reinforce your credibility. The more you’re referenced by trustworthy sources, the more likely search engines and AI models will view you as an authority.
   
   These citations are basically a byproduct of creating original content and building a brand.
   
   For example, dozens of people discuss Exploding Topics on Reddit each month.

   

In the future, you won’t worry as much about ranking #1 in Google.

The real goal will be ensuring your content is the one being cited in AI-generated answers, like this:

Being seen as an authority in your niche will be a BIG part of “ranking” in AI results.

2. Don’t Sweat Technical SEO “Errors”

Technical SEO is important—but most people overdo it.

They obsess over tiny issues that don’t actually impact rankings, spending hours chasing "errors" that Google either doesn’t care about or can work around on its own.

Yes, your site should be fast and crawlable.

But taking your page speed from 1.2 seconds to 1.19 seconds? That’s not going to help you rank.

Technical SEO Has Been Losing Importance for Years

Even before AI started shaking things up, Google was already getting better at dealing with messy websites.

For example… Sitemaps used to be critical.

Today, Google barely relies on them because it can discover and index content just fine on its own.

No, not direct anyway

— Gary 鯨理／경리 Illyes (so official, trust me) (@methode) June 14, 2019

Other examples?

• Messy URLs – Google can figure out URL structures even when they’re poorly formatted.
• Broken internal links – If a page has strong external links or is simply great content, Google still indexes it.
• JavaScript-heavy sites – Google now renders and indexes JavaScript much more effectively than before.

In short, Google has evolved. It’s not as fragile as it once was. And AI-driven search engines will only make technical SEO even less of a priority.

What Actually Matters in Technical SEO Today

Instead of getting stuck fixing minor issues, focus on the few things that actually impact your rankings:

1. Crawlability & Indexability: If Google (or an AI-powered search engine) can’t access your content, you don’t exist. Make sure robots.txt isn’t blocking important pages and that your internal linking structure makes sense. So make sure to still check out the “Indexing” section of GSC on the regular:

   

2. Site Speed (to a Point): If your site loads quickly for humans, that’s good enough. Chasing marginal speed improvements isn’t a good use of time.
3. Mobile Usability: Google is mobile-first. If your site isn’t fast and functional on a phone, that’s a real issue.
4. User experience: Users need to LOVE the experience of reading and navigating around your site. This will never change.

Everything else? Not worth obsessing over.

Where to Invest Instead

If you're spending hours tracking down technical SEO "errors" that aren’t affecting rankings, you’re wasting time.

That energy is better spent on:

• Creating original content that both AI and search engines will cite.
• Earning backlinks that establish your authority.
• Researching your audience to find the next big search trends before competitors do.

Technical SEO is like plumbing—it needs to work, but no one cares how fancy it is.

Fix the basics, then move on to what actually grows your traffic.

3. Topics and Audiences > Keywords

For years, SEO has been all about keywords.

Fire up a tool, find some high-volume search terms, and build your content around them.

That strategy still works—for now.

But in the future, “keywords” are going to be replaced with prompts.

Search Is Moving Beyond Keywords

Google has been shifting away from exact-match keywords for years. Instead, it focuses on meaning and intent.

And AI-driven search engines? They don’t use keywords at all.

A study by AI search by Semrush found that:

• The average Google search query is 4.2 words long.
• The average ChatGPT prompt is 23 words long.
• 70% of ChatGPT queries have “unknown” search intent. Which means they don’t fit traditional search intent categories (informational, transactional, etc.).

This is a HUGE shift.

People don’t interact with AI chatbots like they do with Google. Instead of typing “best running shoes 2025,” they ask:

"I run five miles a day on pavement. I have mild overpronation. What are the best running shoes for support and durability?"

Completely different format. Completely different intent.

How to Research Topics in the AI Era

If you’re still doing SEO the old way, you’ll struggle to rank in AI-driven search results.

Instead of starting with keywords, start with topics and audience research.

Here’s how:

• Talk to your customers. Run surveys, customer interviews, or even casual conversations. Ask them: What are you struggling with? What do you Google when you need help?
• Mine community discussions. Look at Reddit, Quora, Facebook Groups, Slack communities—anywhere people discuss your niche. Pay attention to the language they use and common questions they ask.
• Analyze support tickets and sales calls. Your customer service and sales teams have goldmine-level insights into the real problems people are trying to solve. For example, TK
• Use AI for brainstorming. Instead of a keyword tool, ask ChatGPT: “What are the biggest challenges [your audience] faces?” This will surface pain points, trends, and content ideas you won’t find in keyword research tools.

  

• Create content that AI wants to cite. In the AI era, citations will be the new rankings. If you create high-authority, in-depth pages on important topics, AI chatbots will pull from your content when answering user questions.

For example, Exploding Topics shows up as a source in quite a few LLM results:

4. Scale Back Pure Informational Content

For years, informational content has been a go-to SEO play.

You answer common questions, rank for long-tail keywords, and drive organic traffic.

That strategy is dying FAST.

Why Pure Informational Content Is Becoming Obsolete

Google’s AI Overviews and LLMs are killing search traffic for basic informational queries.

Try searching “How to use garlic chives?”—you’re almost guaranteed to get an AI-generated answer at the top of the SERP.

The days of pulling in thousands of clicks for simple, fact-based content are over.

And honestly? It’s not a huge loss.

Pure informational content rarely converts well anyway.

So instead of wasting time on generic info content, shift toward content that AI can’t easily replicate.

What to Do Instead

If you’re still producing informational content, make sure it:

• Provides unique insights and deep analysis. At Exploding Topics, we still publish informational content. But instead of surface-level answers, we break down trends before they take off with real-world data.

  

  AI struggles with this level of original insight.

• Focuses on mid- and bottom-funnel topics. AI can generate a list of “best CRMs for dentists,” but it can’t actually test those CRMs. If you publish in-depth reviews, pros/cons, and real-world comparisons, you offer something AI-generated content can’t.
• Includes firsthand experience. Whether it’s a product review, industry research, or personal case study, AI can’t replace firsthand testing, interviews, or real-world experimentation.

  

• Brings together multiple sources. AI is trained on existing content, but it struggles to synthesize information from different, high-quality sources into something truly valuable.

Future-Proofing Your Content Strategy

If you’re still pumping out basic, Wikipedia-style articles, it’s time to stop.

Instead, shift toward content AI struggles to create—deep analysis, original insights, firsthand experience, and content that helps users make real decisions.

5. Diversify Your Traffic Sources

If your business relies 100% on Google traffic, you’re playing with fire.

Even before AI Overviews and ChatGPT, depending on Google was risky.

Now? It’s straight-up dangerous.

No one knows exactly where search is headed.

Maybe Google regains trust and stays dominant.

Maybe AI-powered search engines like Perplexity start eating into its market share.

Maybe people skip search entirely and go straight to chatbots or social platforms.

The only safe bet? Diversify your traffic—now.

Why Diversification Is a No-Brainer

Search isn’t the only traffic source shrinking. Google’s organic visibility has been declining for years thanks to:

• More ads pushing organic results further down
• SERP features (AI Overviews, People Also Ask, Featured Snippets) stealing clicks
• Zero-click searches where users get their answer without leaving Google

Meanwhile, alternative platforms are growing—fast.

Social media, newsletters, and free tools are becoming the main discovery channels for many consumers.

At Exploding Topics, we’ve always prioritized building multiple traffic sources so we’re never at the mercy of a single platform change.

Here’s how we do it—and how you can too.

How to Reduce Your Dependence on Google

Here are a few proven ways to drive traffic outside of search:

• Launch a free tool. Exploding Topics offers a free version of our trend discovery tool that anyone can use—no login required.

  

  It’s been a huge source of traffic, especially early on when we had no SEO authority.

• Grow an email list. Our newsletter reaches over 100,000 people with a 40%+ open rate.

  

  Email is algorithm-proof—once someone subscribes, we can reach them anytime.

• Leverage YouTube. Search isn’t just text-based. Our YouTube channel helps us expand our reach beyond traditional search results.

  

• Go all-in on X (or your platform of choice). We post new trends and insights daily on X, building an audience and driving traffic outside of Google.

  

The best platform for you depends on your audience. If you're in B2B, LinkedIn might be a goldmine.

If you're a DTC or lifestyle brand, Pinterest, Instagram, or TikTok could be huge. The key is to experiment, analyze, and double down on what works.

Conclusion

As someone who’s been in the SEO world for over a decade, I’ve had to adapt to countless changes (Panda, Penguin, mobile, voice search, Featured Snippets and more).

And today is no different.

As I said at the jump: AI is a short-term opportunity. But a long-term threat.

I really believe that.

But to take advantage of the short-term opportunity, you need to adapt.

And to sidestep the long-term thread, you need to adapt even MORE.

If you do, you’ll be in a GREAT spot as search evolves.