SEO Is Not Dying. It Is Being Rebuilt From Scratch

AI is not replacing SEO. It is changing what SEO is optimising for. The signals that drove rankings for the past two decades, keyword density, backlink volume, domain authority as a proxy for trust, are being supplemented by something harder to manufacture: demonstrated expertise, clear answers, and content that earns citation rather than just clicks. The discipline is not disappearing. It is getting more demanding.

That shift has real consequences for how marketing teams allocate time, budget, and creative energy. Some of what worked before still works. Some of it is becoming dead weight. The teams that will come out ahead are the ones willing to look at their SEO programme honestly, not through the lens of what they have already built.

What Is Actually Changing in SEO Right Now?

I have been around long enough to remember when SEO meant stuffing keywords into alt text and buying links from directories that existed purely for that purpose. The industry has matured enormously since then, but it has also developed a habit of calling every algorithm update a revolution. Most of them were not. This one is different, and I say that with some reluctance, because I have been burned before by overstating change.

What is genuinely different now is the nature of the search result itself. Google’s AI Overviews, Bing’s Copilot integration, and the rise of conversational AI as a first-stop research tool mean that a significant portion of queries are being answered without a click ever happening. The question is no longer just “can I rank for this?” but “will my content be cited when the answer is generated?”

Those are different problems requiring different solutions. Ranking for a keyword requires satisfying algorithmic signals. Being cited in an AI-generated answer requires being the clearest, most credible, most directly useful source on a topic. One is a game of signals. The other is a game of substance.

If you want a broader view of how AI is reshaping the entire marketing stack, not just search, the AI Marketing hub covers the landscape in more depth.

Has Keyword Research Become Less Useful?

Not less useful. Less sufficient. Keyword research used to be the foundation of an SEO strategy. Map the terms, build the pages, earn the links, watch the rankings move. That workflow still has value, but it is no longer the whole picture.

The problem is that AI-powered search is increasingly interpreting intent rather than matching strings. A user typing “best time to post on LinkedIn” is not necessarily looking for a blog post with that exact phrase in the H1. They may be looking for a direct answer, a tool, a case study, or a nuanced explanation of why the question itself is flawed. AI systems are getting better at inferring which of those the user actually wants.

When I was running paid search at scale, managing hundreds of millions in spend across multiple markets, one of the things that separated good accounts from great ones was the ability to read intent rather than just match keywords. A click on “cheap flights to Barcelona” and a click on “flights to Barcelona” look similar on paper. The conversion behaviour was often completely different. SEO is now facing the same maturity curve that paid search went through a decade ago.

The practical implication is that keyword research needs to be paired with intent mapping. What is the user actually trying to accomplish? What format best serves that need? What would make your answer the one worth citing? Tools like Semrush’s AI SEO guidance and Moz’s breakdown of AI tools for SEO are useful starting points for thinking through how to adapt keyword workflows for this environment.

Does Content Quality Now Mean Something More Specific?

“Content quality” has been a phrase in SEO for years, but it was often used as a vague aspiration rather than a measurable standard. Write good content. Make it useful. Add value. Fine. But what does that mean in practice when an AI system is deciding whether to cite your page?

It means something fairly specific. First, the content needs to demonstrate that the author has genuine knowledge of the subject, not just familiarity with what other pages have written about it. Second, it needs to answer the question directly, without burying the answer behind three paragraphs of preamble. Third, it needs to be structured in a way that a machine can parse cleanly, which in practice means clear headings, logical flow, and concise paragraphs.

I judged the Effie Awards for several years, and one of the things that became clear quickly was the difference between work that looked impressive and work that actually did something. The same distinction applies here. Content that performs well in AI-driven search is not always the most elaborate or the most comprehensive. It is often the most direct. The piece that answers the question cleanly, without padding, without hedging, without the kind of throat-clearing that fills out word counts but adds nothing.

Moz has been doing interesting work on this with their AI content brief tools, which help structure content around the signals that matter rather than just volume. Worth looking at if you are trying to tighten your content production process.

Is Link Building Still Worth the Investment?

Yes, but with a sharper filter on what counts. The era of link volume as a proxy for authority is effectively over. Google has been moving in this direction for years, and AI-driven search accelerates it. A handful of genuinely earned links from credible, relevant sources is worth more than a hundred links from sites that exist primarily to distribute links.

The more useful frame is topical authority rather than link authority in isolation. If your site consistently publishes credible, well-structured content on a defined topic area, and earns citations and links from other credible sources in that space, you build a signal that is much harder to game and much more durable than a link-building campaign built on outreach volume.

Early in my career, I taught myself to code because the alternative was waiting for someone else to give me permission to build something. That same instinct applies here. Teams that build genuine topical authority because they actually know their subject will outperform teams that are trying to reverse-engineer citation patterns. The shortcut is not the shortcut.

The Ahrefs AI SEO webinar with Patrick covers the link authority question in practical terms, including how to think about link quality in an AI-influenced ranking environment.

How Is Search Behaviour Itself Changing?

Search is fragmenting. That is the most important structural shift, and it is the one that most SEO strategies have not yet fully accounted for.

A meaningful portion of search behaviour is moving to AI chat interfaces. Users who would previously have typed a query into Google are now asking ChatGPT, Perplexity, or Claude. These systems do not return a list of ten blue links. They generate an answer, sometimes with citations, sometimes without. The implication for SEO is that optimising purely for Google rankings captures a shrinking share of the total search opportunity.

At the same time, voice search continues to grow, particularly for local and transactional queries. The format of a voice answer is different from a text result. It is typically a single sentence or short paragraph, drawn from a source that answers the question directly and authoritatively.

I remember when paid search felt like a revelation. I launched a campaign for a music festival at lastminute.com and watched six figures of revenue come in within roughly a day from what was, in hindsight, a relatively simple campaign. The reason it worked was not sophistication. It was that we were showing up exactly where the intent was, in the right format, at the right moment. That principle has not changed. What has changed is the number of places where intent now lives.

A strong SEO strategy in 2025 and beyond needs to account for traditional search results, AI overview citations, conversational AI interfaces, and voice. That is a more complex surface area than most teams are currently managing.

What Should Technical SEO Teams Be Focused On?

Technical SEO is not going away, but its relative importance in the overall mix has shifted. A technically clean site is now table stakes, not a competitive advantage. If your site is slow, poorly structured, or difficult for crawlers to parse, you will be penalised. But fixing those issues will not, by itself, move the needle in a meaningful way. The ceiling is lower than it used to be.

Where technical SEO still has genuine leverage is in structured data. Schema markup helps search engines and AI systems understand what your content is about, who produced it, and how it relates to other entities. Getting this right is increasingly important as AI-driven search relies more heavily on structured signals to generate accurate answers.

Page experience signals, Core Web Vitals in particular, remain relevant. Not because they are a dominant ranking factor, but because they are a signal of overall site quality that correlates with the kind of content investment that does drive ranking. Sites that perform well technically tend to be sites that are well-maintained and well-resourced. The signal and the underlying quality tend to move together.

The Ahrefs AI tools webinar has useful material on how to prioritise technical work in the context of an AI-influenced SEO environment, including where automation can help and where human judgement is still required.

How Should SEO and Content Teams Work Together Differently?

This is where I see the most dysfunction in practice. SEO and content have historically operated as adjacent but separate functions. SEO identifies the keywords and the opportunity. Content produces the pages. The brief goes one way, the output comes back, and the relationship ends there.

That model does not work well in an environment where content quality is a primary ranking signal. The SEO team needs to understand what makes content genuinely authoritative on a topic, not just what keywords it should contain. The content team needs to understand the technical and structural requirements that make content eligible for AI citation. Neither team can do their job well without the other.

When I grew an agency from 20 to 100 people, one of the things that consistently created problems was functional silos that made sense on an org chart but produced poor work in practice. The best results came when people with different specialisms were working on the same problem rather than passing work between departments. SEO and content in 2025 is exactly that kind of problem.

Practically, this means content briefs need to be richer than a keyword list and a word count target. They need to include the specific questions the content is trying to answer, the expertise signals it needs to demonstrate, the sources it should reference, and the format that best serves the user’s intent. HubSpot’s work on AI marketing automation covers some of the workflow implications here, including how AI tools can help bridge the gap between SEO insight and content execution.

What Does Good SEO Measurement Look Like Now?

Organic traffic and keyword rankings are still worth tracking, but they are increasingly incomplete as measures of SEO performance. A page that is cited in an AI Overview may drive significant brand awareness and downstream conversion without generating a single direct click. A ranking that appears below an AI-generated answer may see its click-through rate decline even as the ranking itself holds steady.

I have always been sceptical of measurement systems that mistake the proxy for the thing. Analytics tools are a perspective on reality, not reality itself. In SEO, that means building a measurement framework that captures the full picture: rankings, yes, but also AI Overview appearances, branded search volume as a proxy for awareness, direct traffic trends, and conversion rates from organic sources over time.

Share of voice across both traditional search and AI-generated results is becoming a more meaningful metric than position alone. If your content is being cited by AI systems as the authoritative answer on a topic, that has commercial value even when it does not show up cleanly in a rankings report.

There is more on the broader AI marketing measurement question across the AI Marketing hub, including how to think about attribution in an environment where the path from search to conversion is increasingly non-linear.

What Is the Realistic Timeline for These Changes?

Longer than the hype suggests, and shorter than the sceptics would have you believe. That is almost always the answer with structural shifts in digital marketing.

AI Overviews are already live and already affecting click-through rates on informational queries. Conversational AI as a search interface is already a significant behaviour for a meaningful segment of users, particularly younger demographics and technically sophisticated audiences. These are not future trends. They are current conditions.

At the same time, traditional search results are not going away. The majority of search volume still flows through conventional results pages. Most users still click on links. Most conversions still start with a keyword search in a conventional sense. The transition is real but gradual, and the teams that panic and abandon proven fundamentals will be worse off than the teams that evolve methodically.

The practical advice is to run both tracks in parallel. Maintain the fundamentals that still work. Build the capabilities that the new environment requires. Do not wait until the shift is complete to start adapting, because by then the competitive gap will already be established.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions