Yes, You Can Do SEO With AI Now. That's Exactly Why Most of It Will Fail.

Growth Marketing Sweden is a finalist for Best Use of AI at the Global Search Awards 2026. Founder Jorge Castro on the most dangerous misconception in search marketing right now — and why the signals that decide rankings today were never inside the content to begin with.

Here’s a conversation we have almost every week now.

A business leader tells us they’ve started doing SEO with AI. They ask ChatGPT or Claude what to do, they get sensible advice, they produce good content, and the titles are optimised. Everything looks right.

And nothing happens.

Not because the AI was wrong. Because it was generically right — and generically right is precisely what doesn’t rank anymore.

The advice isn’t wrong. It’s just not a strategy.

Ask an LLM how to do SEO, and you’ll get a competent summary of best practice: write quality content, match search intent, optimise your titles, build internal links. All correct. All published a thousand times. All available to every one of your competitors in the same three seconds it took you to get it.

That’s the problem. When everyone has instant access to the same “best practice,” best practice stops being an advantage. It becomes the entry fee.

What the AI can’t see is everything that actually decides the outcome: your existing content and how it performs, what your Google Ads data says about which messages convert, your backlink profile compared to the competitors above you, whether anyone outside your own website ever mentions your brand. An LLM gives advice in a vacuum. SEO is won in context.

Google stopped rewarding “good title, good text” a long time ago

This is the part that surprises people most: none of this is new.

Many still picture SEO as it looked fifteen years ago — a well-written page, the right keywords, a decent title tag. But Google has spent over a decade systematically moving the goalposts away from what’s written on the page and toward signals that are earned off it.

When Google formalised E-A-T — Expertise, Authoritativeness, Trustworthiness — in its quality guidelines back in 2014, it was making something explicit that had been building for years before that: Google doesn’t just evaluate what you say. It evaluates whether the rest of the internet gives it any reason to believe you.

Authority is not a paragraph you write. It’s the sum of signals you earn: who links to you, who mentions you, whether journalists cite you, whether people search for your brand by name, whether your expertise shows up anywhere other than your own blog. Adding “Experience” to the framework in 2022 only sharpened the point — Google wants evidence of real humans with real experience, verified by the world around them.

And in the era of AI-generated content, those external signals matter more than ever, not less. When anyone can produce unlimited plausible text, the text itself proves nothing. Trust has to come from somewhere the machine can’t fabricate. That’s also exactly how the new AI search experiences choose what to cite: brands that the wider web already treats as authorities.

SEO is not a channel. It’s the scoreboard for all your channels.

This is why we tell clients something that sounds strange coming from a search agency: a large part of modern SEO doesn’t happen on your website.

It happens in digital PR that earns the mentions and links Google reads as trust. It happens in your Google Ads account, where real conversion data reveals what your audience actually responds to — intelligence most “AI SEO” never touches. It happens in the content you already have, which often holds more ranking potential than anything new you could write. And it happens in your brand, because branded search demand is one of the strongest signals there is.

Content is one instrument. SEO is the orchestra.

So where does AI fit? Everywhere — one level down.

This is what our shortlisting for Best Use of AI actually recognises. Not that we let AI do our SEO — that would be like letting the calculator do the accounting. But used inside a real strategy, AI is transformative: it analyses search intent, competitors and content gaps at a scale no human team could match, it accelerates production once the strategic decisions are made, and it frees our strategists to spend their time on the work that moves rankings — the signals, the PR, the context.

Humans decide. AI accelerates. And the strategy has to see the whole board.

If an agency — or a chatbot — proposes an SEO plan without asking about your existing content, your ads data, your link profile and your PR, it isn’t proposing a strategy. It’s reciting one.

We’ve seen how this movie ends

We know, because we keep meeting the sequel.

Time and again, companies come to us after running their SEO through an LLM. The pattern is always the same: at first, a beautiful graph. Content goes out at a pace they’ve never managed before, impressions climb, and it feels like they’ve cracked the code. Then the curve bends — and everything goes south. Rankings slide, traffic evaporates, and sometimes the site ends up in a worse position than before it started.

Scroll LinkedIn any given week and you’ll find another post-mortem that follows this exact script. This is not bad luck. It’s Google doing what Google has done for fifteen years: identifying the newest shortcut at scale and closing the door on it. Mass-produced, context-free content is simply the latest shortcut — and Google has been remarkably efficient at shutting it down.

Meanwhile, look at where the smart money is going. Adobe didn’t just dabble in SEO — it acquired Semrush for 1.9 billion dollars to build search visibility into the core of its marketing stack. And the global SEO services market is projected to grow at over 13% a year, reaching an estimated 171 billion dollars by 2030. The world’s largest software companies are not betting billions on a discipline that a chatbot just replaced. They’re betting on it because search — in all its new forms — is becoming more strategic, not less.

The bottom line — and what we’re building

Yes, you can do SEO with AI now. So can everyone else. Which means the winners won’t be the ones with the best prompts — they’ll be the ones who understand that Google made this a multi-signal game long before AI arrived, and who use AI to play that game faster.

That’s the philosophy behind jorgeCastro.AI, the platform we’re building and the work recognised at the Global Search Awards. Our goal is SEO generated from what actually decides rankings: key data points, live connections, APIs, MCP integrations and the full context of a client’s channels — orchestrated across multiple LLMs, each doing what it does best. It’s why a single piece of content can take up to 20 minutes to produce instead of three seconds. Those extra minutes are the strategy.

Fast text is everywhere now. Earned trust isn’t. We know which one Google rewards — because we’ve watched it close the door on everything else.