AI did not kill SEO. It changed where the clicks go and which pages deserve them. Here is what actually shifted, which pages lost traffic, and the five places AI genuinely helps the work.
The question gets asked as though search is about to switch off. It is not. What has happened is narrower and more useful to understand: AI summaries sit at the top of more results, they answer a certain kind of question directly, and the traffic that used to flow from that kind of question has largely stopped flowing.
If your traffic came from those queries, it feels exactly like AI killing SEO. If your traffic came from people deciding whether to spend money, very little has changed, because nobody chooses a contractor, a clinic or a software platform from a summary paragraph.
Three shifts, in rough order of impact.
Informational queries resolve on the page. What is a sitemap, how long does grout take to dry, what does a tooth crown cost on average. These produce an answer box and no click. Pages built around that kind of query have lost the most.
The results page got taller. Even where a click is still likely, there is more above the first organic result than there used to be. Position three now behaves more like position six did.
Search behaviour got longer. People type conversational, specific, multi part queries now because they have learned that machines can handle them. That favours pages that answer a real question properly over pages built around a two word keyword.
Pages that have lost the most tend to share a profile. Broad definitional topics with no commercial angle. Content written to occupy a keyword rather than to say something. Pages that could be swapped with a competitor's page without anyone noticing. Thin location pages differing only by town name.
None of that is new advice. AI just removed the margin that let those pages survive.
Notice what is missing from that list. The experience, the opinion, the numbers from your own account, the thing you learned doing the job. That is the part that makes a page worth ranking, and it is the part AI cannot supply, because it does not have your data.
Audit which of your pages lost traffic and split them into two piles. Pages that lost informational traffic that was never going to convert can be left alone or consolidated. Pages that lost commercial traffic have a different problem, and it is usually that a competitor made a better page.
Then make the remaining pages harder to replicate. Put your own numbers in them. Name the people with the expertise. Show the work rather than describing it. Answer the specific question in the heading in the first two sentences underneath it.
That approach ranks conventionally, it gets cited in AI answers, and it survives the next shift too, because it is not a trick.
Many businesses make the mistake of treating AI-generated content as a complete replacement for strategy, expertise, and human oversight.
Keyword research and topical organization can take hours.
AI tools can dramatically speed up this process.
You can ask tools like ChatGPT to:
Afterward, validate those keywords using SEO tools like:
AI is incredibly useful for building the foundation of a webpage.
For example, you can ask AI to generate:
This helps writers structure content more comprehensively around a target keyword.
AI can also help create important on-page SEO components, including:
For example, you can prompt AI to:
Schema markup can be time-consuming to write manually.
AI tools can quickly generate structured data formats like:
This saves time researching syntax and implementation.
AI can also assist with SEO analysis and reporting.
For example, marketers can:
This helps teams make faster, data-driven SEO decisions.
No, but it has removed a large share of informational clicks. Queries with one clear factual answer now get resolved on the results page. Commercial, comparison and local queries still produce clicks, because people spending money want to see the provider rather than a summary. The pages that lost most were the ones written to occupy a keyword rather than to say something useful.
It replaces the parts of the job that were mechanical: clustering keywords, drafting outlines, generating schema, parsing crawl data. It does not replace judgement about what a business should target, what makes a page worth ranking, or how to turn firsthand expertise into something a search engine can see. The role shifts toward strategy and evidence.
Google's position is that it judges the content, not how it was produced. In practice AI drafts rank badly when they are generic, because that is what generic content does. AI output used as a starting point and then loaded with real experience, original data and specifics tends to perform the same as anything else written to that standard.
The pages being cited are overwhelmingly pages that already rank well conventionally. Clear structure, direct answers placed immediately under the heading that asks the question, accurate structured data and genuine topical depth are what make a page easy to quote. There is no separate AI optimisation channel to buy into.
It is good for the research and production scaffolding: grouping topics, drafting outlines, writing metadata at volume, summarising data exports. It is poor at supplying substance, because it has no access to your results, your customers or your experience. Use it to go faster at the parts that are not the differentiator.
Broad definitional and how-does-it-work pages with no commercial angle, thin location pages that differ only by town name, and interchangeable keyword pages. Pages tied to a purchase decision, a local service or a genuinely specific problem have held up considerably better.
Our in-house team works directly with US-based marketers. Tell us what you are trying to grow and we will tell you what we would do first.
Drop in your website and we will record a short teardown: where you are losing traffic, who is outranking you, and the three things we would fix first.
A real person records it. No obligation, and we will only use your email to send the video and talk about it.