From SEO to AIO

In the beginning there were meta tags. Webmasters (I see you shudder) filled hidden HTML tags on web pages with keywords. Search engines, working from CEOs’ parents’ garages, trusted webmasters and placed their sites accordingly. But webmasters weren’t to be trusted, and the SEO arms race was born. Humans tried ever harder to game meta tags, and AskJeeves, AltaVista and then Google evolved to be more discerning.

The next phase of evolution saw search engines focusing where they really should have focused all along – on human-readable content. Writing about your company’s new ISO-compliant payroll consultancy while filling your meta tags with Britney Spears-focused keywords (I was there. This was the early 90s. It really did happen like this) no longer worked. For the cynical, writing content in order to be found became:

• Put your keywords in the title
• Put your keywords in the first paragraph
• Put your keywords in subheadings and lists
• Put your keywords in the alt text
• Tattoo your keywords on your partner’s arm

…and if you did that enough times, Google might reward you with visitors – or existential validation.

As search engines matured further, they moved onto more behaviour-driven algorithms, still with content and key phrases at the centre, but with a much better understanding of how users actually interacted with your content – or indeed whether they did. Throughout, classic SEO essentially felt like a polite (ish) negotiation with a robot. You could feel the complexity of the algorithm, but it still definitely felt like an algorithm.

You… “Here is a page about heritage venue hire in Bath.”
Google… “I see you have said ‘heritage venue hire in Bath’, or similar language, around fourteen times, and my data shows that users are spending some time on your page. I’ll try to place you a little higher and see what happens.”

The content had to engage humans to some extent, but it also had to be structurally obedient, often slightly repetitive and written for humans, but with web crawlers in mind. If you wanted to maximise your advantage or simply game the system, these were simpler times. Not quite as simple as choosing a company name beginning with A to appear higher in the Yellow Pages, but certainly simpler than today.

We are now, in case you missed it, in the early phases of a much bigger revolution… a revolution we are only just beginning to understand the scale of and begin to adapt to in terms of SEO. AI via LLMs has arrived and has started answering questions directly, and, annoyingly for those still gaming the system, very informatively.

The robot you were trying to win over now has a gatekeeper – a much, much smarter robot.

The New Reality: AI as the First Reader

Today, your content has three audiences:

• Humans
• Search engines
• AI systems that summarise you before humans ever see you

Increasingly, people don’t click through to your article or your product, and they may not even see you in their search results. With luck, you might be one of the sources your friendly neighbourhood LLM refers to. More often than not, you might not be.

AI systems don’t just count keywords. They don’t even just measure behaviour. They extract meaning. They compress your ideas. They compare you to others. They decide whether you’re clear, authoritative, derivative, confused, or quietly brilliant. This means that some of the old tricks don’t just fail – they may actively backfire.

Keyword stuffing in the age of AI is like shouting to be understood in a foreign country.

The AI vs SEO Transition

The awkward truth is that an awful lot of easy-to-find content from the last decade was written to game a system (Google any recipe if you want abundant evidence), and AI was trained on that content. So we now have systems trained on entire oceans of: “Top 10 tips for making jam roly-poly,” “The ultimate guide to goat husbandry,” and “Everything you need to know about craft breweries in Basingstoke.”

The result? AI answers are in danger of sounding like… well… all of that same, SEO-focused, system-gaming content. If everyone writes generic SEO copy, AI will summarise generic SEO copy. But – and here’s the gap – if you write something distinctive, specific, and thoughtful, AI has something better to learn from… and to cite.

Dare we whisper it – it might actually, finally, be about quality now.

So, if we’re not keyword stuffing, can you just tell me what to do?

Absolutely – that’s the point. In an AI-shaped search landscape, the incentives shift.

Depth, not density

It’s not about repeating keywords and phrases; it’s about answering the question better than anyone else.

Originality, not reformulation

If your article is a rephrasing of the first five search results, AI doesn’t need you. But if your article contains a perspective, data, synthesis, or lived experience, you become useful.

Clarity not cleverness

AI systems reward structured, well-explained ideas. Clear headings. Logical flow. Real substance. (Sorry AI, I should have tried harder.)

More than ever, good writing is the best SEO strategy. AI-oriented SEO, let’s call try calling it AIO until we realise it will never take off, is forming around a different question: how do we become a trusted source?

The good news is that, for organisations built on knowledge – arts and cultural organisations, membership bodies and often charities – this shift is less of a threat and more an invitation. You’re not filling pages of content about top plumbing tips in order to sell shower heads – you already have the expertise. The key is how to engage the experts in your organisation and how to get that expertise onto your site.

Five Tips for Writing AI-Focused Content (see what we did there?)

• Publish genuinely helpful content
• Develop and show recognisable expertise
• Have a coherent, in-depth point of view
• Say something worth quoting
• Express your personality

How Do We Get Our Expert Content Out There? A Health Warning

Of course, the arms race isn’t over. There will be a flood of AI-generated content optimised for AI search. It will be:

• Grammatically flawless
• Mildly informative
• Completely forgettable
• Full of lists
• Full of telltale signs — for example, the em-dash

The web may become a vast plain of perfectly competent but clearly repetitive beige.

So what’s the answer if you want to stand out?

• Human specificity
• Use examples, stories, judgement and taste.
• Don’t be scared of slightly unfashionable opinions or occasional humour.
• Don’t be afraid to have personality.

AI can synthesise the average, but it still struggles to replace the distinctive. And please – stop putting the keyword in the subheading fourteen times.