The arrival of AI has changed search like nothing before it. More and more people look for information in ChatGPT, Perplexity, Claude or Gemini without ever running a traditional search. The blue links we all grew up with are losing ground to answers generated on the spot.
Does SEO still make sense when more and more people ask AI instead of googling? Are we watching the end of the industry as we know it?
SEO has been evolving since day one, toward the same goal — being the best possible answer to a question.
Let’s look at it with some perspective.
Google’s Success and the Birth of SEO
Before looking forward, you have to understand what came before. Google didn’t win the search engine wars by accident — it won because it built an algorithm that answered a basic human need better than anyone else: finding information fast and reliably.
To do that, it introduced PageRank in the late 90s. Until then you had hand-built directories like Yahoo! and engines that ranked by keyword matching, like AltaVista. Brin and Page’s idea was to look at links — and not just count them, but weight them by who was linking. A link from a trusted source was worth more than a hundred from anywhere else.
Google started paying attention not only to what a page said about itself, but to what everyone else said about it.
In that context SEO was born, with two faces from the very first day. On one side, understanding the algorithm: what Google measures, how it measures it, and what you need to give it so it understands you. On the other, building enough authority for Google to take your site seriously. In fact, the 1998 paper (opens in a new tab) itself already saw it coming: Brin and Page noted the existence of “numerous companies which specialize in manipulating search engines for profit”.
For more than 25 years, Google has owned web traffic. Around 90% of global searches (opens in a new tab) go through its engine. It has shaped what gets read, what gets bought, what gets seen. And it has changed the rules along the way.
With Google distributing the traffic, position one became the jackpot. And keeping the algorithm happy became priority number one for anyone who wanted visibility online.
Google was well aware of this. It kept refining its algorithm with the same goal it always had: give the best possible answer to every query so it stays the platform everyone uses. Every time someone found a shortcut, Google closed the door.
These are the updates that defined an era:
| Year | Update | What changed |
|---|---|---|
| 2011 | Panda (opens in a new tab) | Penalised low-value sites with copied or unhelpful content, and rewarded original content backed by real research and analysis |
| 2012 | Penguin (opens in a new tab) | Went after link spam. Since 2016 it's part of the core algorithm and runs in real time |
| 2015 | Mobile-friendly update (opens in a new tab) | Boosted mobile-adapted pages in mobile searches. Applied page by page, across all languages |
| 2018 | August broad core update (opens in a new tab) | A broad core change. Hit health, finance and legal disproportionately. Google said there was nothing specific to fix |
| 2019 | BERT (opens in a new tab) | Natural language processing to understand search intent. Affected 1 in 10 English searches |
| 2022 | Helpful Content Update (opens in a new tab) | Introduced a site-wide signal against content made mainly for search engines rather than people |
| 2024 | March 2024 Core Update (opens in a new tab) | Folded the helpful content system into the core and added new spam policies. Google estimated a 40% reduction in unhelpful content |
An Unexpected Update: Artificial Intelligence
Google had spent decades updating its algorithm, but always from the inside. In November 2022 something shifted: OpenAI launched ChatGPT. Within two months it passed 100 million users. Today it’s over 900 million every week (opens in a new tab).
For the first time there was a real alternative to “googling it”. Not another competitor, but a different way of asking: direct conversation, a synthesised answer, no need to click anything.
Google responded with Bard in 2023, Gemini after that, and AI Overviews inside its own results from May 2024. Today AI Overviews are past 2.5 billion users a month (opens in a new tab), and AI Mode, launched barely a year ago, is already over a billion.
Search stopped being “I type keywords and click a link”. A layer of AI has been added on top of the same old blue links, also trying to resolve the user’s intent. Sometimes leaning on those links, sometimes not.
Interestingly, in that same announcement Google stated that searches hit an all-time high last quarter.
It doesn’t look like replacement, but evolution.
GEO, AIO, AI SEO — New Labels for Old Work
The last few months have brought an explosion of acronyms: GEO (Generative Engine Optimization), AIO (Artificial Intelligence Optimization), AI SEO, LLMO (Large Language Model Optimization). More will follow.
My take: most of them are a new label for the same old thing. The name itself gives it away — Search Engine Optimization, not Google Optimization.
You could object that an LLM isn’t a search engine, and that’s partly true. But when you ask ChatGPT with search enabled, or Perplexity, or Gemini, the system runs queries, pulls results from an index and synthesises them. The search is still there. What changed is who runs it: before it was the user, now it’s the model on their behalf.
The goal was always organic visibility, and the fact that for more than 25 years this meant Google was a circumstance, not a definition.
Before, only Google mattered in most cases. Now it’s also ChatGPT, Perplexity, Gemini or Claude. But the underlying work hasn’t changed: making them find your content and pick it to answer a query.
The Three Pillars — and What Changes in Each
It doesn’t matter whether we’re talking about ranking on Google in 2015 or showing up in Perplexity in 2027. The work comes down to three things, and none of them are new. As always, when SEO evolves you have to adopt new practices in each area.
Yes, I’m about to simplify SEO work so hard it hurts me too, but I do think these are the three big pillars that stay intact.
Accessibility
What hasn’t changed: if a bot can’t reach your site, you don’t exist to it. No server-side blocks, clean HTML, a robots.txt you’ve actually looked at, and rendering that doesn’t get in the way of extracting your content.
What’s new: it’s not one bot anymore, it’s several. Googlebot, GPTBot, ClaudeBot, PerplexityBot… Plenty of sites are blocking AI bots without knowing it, or because someone decided so at the time without weighing the consequences.
Content Quality
What hasn’t changed: answering the question better than anyone else. Original content, clear structure, concrete examples, verifiable data, and a decent experience while reading it.
What’s new: with AI generating text at scale and at a fairly acceptable quality, the bar goes up.
What sets a site apart now is the real experience of the person behind the business.
I think this is where most sites fail. Often there’s no real connection between what’s on the site and the person running the business. If AI writes it, or an SEO writes it with nothing else to go on, you always leave out the part that actually makes the difference. Easy to say, hard to find people who go page by page making sure it matches what the business actually offers.
Authority
What hasn’t changed: what other people say about you — the same idea we started with, PageRank. For years that was measured almost entirely through backlinks.
What’s new: links lose relative weight against other signals:
- Mentions in the outlets that matter in your industry
- Consistent presence in trusted sources: LinkedIn, Wikipedia, press
- Verifiable authorship: who you are, what you’ve done, where you’ve worked
- Brand reputation: if people search your name, that’s real authority
Accessibility, content and authority reinforce each other. Failing at one can bring the whole thing down.
Measurement Is Where the Real Shift Is Happening
For many years the metrics were the same: position and impressions gave you visibility, clicks gave you organic traffic. With AI it gets more complicated. You have to account for citations, with and without a link, and often there’s no click at all because the user just takes the answer.
We do have some tools already — Bing Webmaster Tools, a bit inside Search Console, purpose-built tools like Peec.ai or Otterly.ai — but I expect Google to move fast here.
Search Console is where SEOs have looked first thing every morning for twenty years, and if that conversation moves to third-party tools, Google loses more than data. I’d expect AI visibility reports inside Search Console sooner rather than later.
Building a Brand Will Be the Deciding Factor
Scrolling through Instagram, it struck me that the future of online business might not be so different from how that platform works. Its algorithm recommends accounts you don’t follow but that match your interests.
I got shown one that breaks down surf forecasts, for instance. Not famous, but the content helps me. I follow it, and that’s one more user for them.
Something similar happens with AI: it gives you a personalised answer and you discover sites you’d never heard of. The difference is that Instagram recommends by content affinity, while AI recommends mostly by trust. It needs to know you exist (accessibility), that you are who you say you are, and that others back it up (content + authority).
In that context, being remembered will matter more and more. AI, like people, tends to cite what it has already seen several times from different sources. Strong brand, community and consistent presence across channels will decide who shows up when someone asks for a “trip planner in Indonesia” or a “surf coach in Lisbon”.
And once they’ve found you, the brand does the second job — the more important one: next time they come straight to you, skipping ChatGPT, where your competitors are fighting over that citation.
So What’s SEO’s Role Going to Be?
An SEO’s job was always to bridge a set of bots and their algorithms with a business’s value proposition. That need doesn’t disappear because the middleman changed.
It used to be Google. Now ChatGPT, Perplexity, Claude and Gemini have joined in, and tomorrow it’ll be others. Even so, the underlying question is the same:
When someone runs a query that’s relevant to your business, why should your site be the one that shows up?
Search engines and language models will always share the same goal: give the user the best possible answer so they can own the market.
That said, I think there will be far fewer profitable websites in the future. The extremes will survive. The giants that win on scale, like Amazon, and the specialists who own a niche — the ones who know their audience better than anyone.
And there’s still work in that. I might be wrong, but I don’t see an AI agent distilling a business’s value proposition better than the person running it, or adapting to an environment that shifts every month. Not in the next three to five years, at least.
That’s how I see the future of SEO. If you want to know where your site stands in all this, get in touch. The first review is free, and you’ll walk away knowing more than you did before — whether we end up working together or not.