Traditional SEO tries to win a position in a list of links. AI SEO tries to be named inside a generated answer. About seventy per cent of the underlying work, technical health, real expertise, third-party credibility, serves both. The thirty per cent that differs is mostly in how you write and how you prove the business is real.
There is a lot of money being raised on the claim that SEO is dead and something new replaces it. That is not what the data shows, and it is not what we see in client accounts. What is true is that a slice of search behaviour has moved, and the writing that wins that slice is shaped differently.
The seventy per cent that is the same
- Technical health. A page that cannot be crawled cannot be ranked or quoted.
- Genuine expertise. Both systems are trying to identify who actually knows the subject.
- Third-party credibility. Links and mentions on trusted sites signal the same thing to both.
- Site structure. Clear hierarchy and internal linking help a crawler and a retriever equally.
- Freshness. Both prefer content that has been maintained.
- Structured data. Schema makes meaning explicit either way.
If your technical foundation is weak, no amount of AI-specific work will help. The retrieval systems are reading the same pages Googlebot reads.
The thirty per cent that genuinely differs
| Traditional SEO | AI SEO | |
|---|---|---|
| Goal | Rank in the top ten | Be named in the answer |
| Unit of value | The page | The sentence |
| Writing shape | Comprehensive, keyword-aware | Answer-first, self-contained claims |
| Winner count | Ten links per query | Two or three names |
| Measurement | Position and clicks in Search Console | Manual prompt tracking |
| Brand mentions | Helpful | Central. Unlinked mentions carry weight |
| Recency | Matters on some queries | Matters on most |
The unit of value changes
Google ranks a page. An assistant lifts a sentence. That single difference drives most of the writing changes: state the claim before the context, keep one idea to a paragraph, use numbers instead of adjectives, and make sure any sentence could be quoted without the paragraph around it.
The winner count changes
Ten organic positions get some traffic. An AI answer names two or three businesses. That is a harsher distribution, and it is why entity clarity matters more here than it does in ordinary SEO. When an assistant is unsure whether two listings are the same business, it names a competitor it is sure about.
Unlinked mentions start counting
Traditional SEO values links. AI systems weigh being talked about, whether or not the mention carries a link. A podcast transcript naming your company, a forum thread recommending you, an industry roundup listing you: all of it is evidence.
Where to put the money
For almost every business today, organic search sends far more traffic than every AI surface combined. Our working split is about eighty per cent traditional and twenty per cent AI-specific, and we would revise it if the numbers moved.
- If your technical base is weak. All of it goes to traditional. Nothing else works until the site can be crawled and rendered properly.
- If you rank well already. This is the best position to be in. AI Overviews draws from the top ten, so a share of AI visibility comes almost for free once the writing is adjusted.
- If you are in a fast-moving category. Weight AI higher. Software, finance and anything where buyers research heavily are further along the shift than local services.
- If you are a local business. Weight traditional higher, but do the entity work. Local AI answers are far less contested than commercial ones, so the entity fixes pay back quickly.
The honest summary
AI SEO is not a new discipline. It is SEO with a different reader in mind: one that reads faster, quotes rather than links, and names fewer winners. The businesses doing well in AI answers are, with very few exceptions, the ones that were already doing SEO properly.
Questions we get asked
Is AI SEO just a rebrand?
Partly, yes. Much of what is sold as AI SEO is ordinary technical and content work with a new label. There is a genuine remainder, mostly entity clarity, citation-shaped writing and third-party mentions, but it is smaller than the marketing suggests.
Should I stop doing traditional SEO?
No. Organic search still sends far more traffic than every AI surface combined for almost every business. AI search is a growing supplement, not a replacement.
Which should I spend on first?
Traditional, if your technical base is weak or you rank poorly. AI Overviews draws from pages already ranking, so ranking is the entry ticket to a good share of AI visibility anyway.
Do the two ever conflict?
Rarely. The main tension is length: AI systems favour tight, quotable passages while some competitive queries still reward depth. The answer is usually a clear summary at the top of a thorough page.
How do I measure AI SEO?
Manually, for now. A fixed prompt set run monthly across ChatGPT, Perplexity and AI Overviews. Search Console does not break out AI Overview impressions.
Is AI search traffic worth the effort?
Per visit, yes: intent is high because the comparison has already been done. In volume, not yet for most businesses. The work is worth doing because it overlaps so heavily with SEO you should be doing anyway.
