Is SEO Dead After AI? The Honest Answer

The question reaches us weekly from clients and site owners alike: if the search engine answers the user directly at the top of the page, why pay for content nobody will click?

The question is legitimate and the anxiety rests on a real observation. The common conclusion drawn from it, however, is wrong. What is happening is not the death of SEO but a harsh redistribution: one category of content lost its value entirely, and another became more valuable than it was.

What actually changed

The search engine used to be an intermediary presenting ten links and leaving the choice to you. It now reads sources, assembles an answer, and presents links as attribution.

The immediate consequence is that queries with a short definitive answer lost their clicks. Somebody looking up an exchange rate, a term's definition, or a simple setup step receives it on the page and visits no website. If your content was built on that category, you have genuinely lost, and you will not get it back.

The second consequence is less obvious: queries tied to a decision — a comparison between two options, choosing a supplier, evaluating a price, an unusual situation — became more valuable per click. The user reads the summary and then looks for a source they trust before spending money. A summary is not sufficient basis for a decision involving money.

Who loses and who gains

Losing: duplicated content restating what ten other sites already say, generic definition pages, and articles written to fill a publishing calendar with no real question behind them.

Gaining: content carrying something a model cannot generate from thin air — data from actual operations, a documented experience, real local pricing, objections heard from customers, detail specific to one market. That is precisely the basis of the work at Naaktob and of Naaktob.

The compressed rule: if a language model could write your article without access to your information, your article did not deserve the click in the first place. That was true before AI; AI simply made the penalty immediate.

Visibility is now broader than ranking

The old metric — "where do I rank" — is no longer sufficient on its own. The question became: is my site cited in the generated answer?

Citation follows logic similar to ranking but stricter: a clear, well-structured source answering a specific question in direct language is cited more often than a long, meandering article. In practice this means:

Lead with the answer. Open each section with the answer in two sentences, then elaborate. Long rhetorical introductions hurt twice: they push the reader away and weaken the chance of citation.

Structure with headings that match real questions. A heading resembling the user's question is easier to capture than a clever, oblique one.

Cite figures, dates and sources. Specific content is cited more than general content, and believed more by the reader as well.

Use structured data. Marking up FAQs, articles and products makes your page's structure machine-readable without ambiguity.

What did not change at all

Three things survived the noise unchanged.

Search intent governs everything. Somebody typing "best web design company" will not buy from an introductory article about the importance of websites. Matching page type to intent type is the core of the work, and it remains entirely human.

Speed and technical experience. A slow page loses the visitor before they read, and a visitor who bounces straight back to the results is a clear negative signal. That is Noprmg work, not a writer's.

Links and mentions. Being referenced by other sources remains the strongest trust signal, and it has grown in importance because models lean on what is widely cited.

What we changed in practice

Three shifts in how we work over the past year.

Fewer pieces, deeper each. Instead of twenty articles a month, materially fewer, longer and more precise, with strong older articles updated rather than weak new ones produced. Updating an article already ranking well returns more than writing a tenth article on a saturated topic.

Decision pages first. Comparison, pricing and case pages take priority over introductory articles, because those are the pages people still click.

Different measurement. We no longer judge by the count of ranking terms alone, but by share of visibility, the proportion of visits with commercial intent, and the number of conversations that began in organic search. Metric selection is covered in how to measure a campaign.

A ninety-day plan for an existing site

If you run a site and have watched traffic slide, this is the order of work.

Weeks one and two: an honest inventory. Sort your pages into three groups — pages bringing commercially valuable traffic, pages bringing traffic of no value, and pages bringing nothing. Most sites discover that under a tenth of their pages produce the entire result.

Weeks three and four: rescue what deserves it. Take the valuable pages and update them: add current figures, rewrite introductions to lead with the answer, add an FAQ section addressing what people genuinely ask, and repair internal links pointing at them.

Month two: merge or delete. Similar pages competing for the same query are merged into one stronger page with a permanent redirect. Pages bringing nothing and serving no reader are deleted without sentiment — a smaller, stronger site beats a larger, bloated one.

Month three: build decision pages. An honest comparison between you and the alternatives, a pricing page explaining the logic even without a final number, and two worked cases with real figures. These are the pages that capture the click a summary did not absorb.

Three signs your SEO supplier is selling illusion

They report ranking only. A report full of keyword positions with no link to valuable traffic or actual conversations is a report designed to look good.

They produce a fixed quantity for no reason. "Eight articles a month" is not a strategy; it is a number chosen to fill a contract. Strategy is knowing why these topics and in what order.

They ask you for nothing. Content that differentiates you needs your knowledge: your numbers, your customers' objections, your operational detail. A supplier who asks for nothing produces content anyone could produce.

How long before results

Largely unchanged: four to eight months for a relatively new site to see organic traffic with commercial value, faster for a site with history. Anyone promising results within a month is selling an illusion, and anyone claiming SEO is dead in order to sell you advertising alone is selling a different one.

The short version

SEO did not die. The content that never deserved to exist did. The opportunity is now better for anyone holding genuine knowledge, because competition on volume lost its point overnight.

To review your existing content through this lens, write to us, or read our sectors.