Two years ago the brief was "get us to page one". Now the same marketing lead asks a different question: why does ChatGPT recommend three competitors and not us? Both questions are about visibility. They are not answered by the same work.
AI SEO and traditional SEO share a foundation and then diverge — on what gets optimised, how success is measured, and where the work actually sits. Here is the honest split, without the pretence that everything you learned is obsolete.
What is AI SEO?
AI SEO is the practice of getting your content cited inside AI-generated answers — Google AI Overviews and AI Mode, ChatGPT, Perplexity, Claude, Copilot — rather than only ranked in a list of blue links. It goes by several names: answer engine optimization (AEO), generative engine optimization (GEO), LLM optimization. The label matters less than the shift underneath it: the unit of visibility moves from the page to the passage.
Confusingly, "AI SEO" is also used to mean using AI tools to do SEO work — drafting briefs, clustering keywords, generating schema. That is a productivity story, not a strategy one. This article is about the first meaning: optimising for AI systems, not with them.
How is AI SEO different from traditional SEO?
Traditional SEO competes for a position in a ranked list; AI SEO competes to be one of the sources an answer is assembled from. Everything else follows from that one change.
| Traditional SEO | AI SEO | |
|---|---|---|
| Unit optimised | The page | The passage |
| Goal | Rank in the top results | Be cited in the answer |
| Query model | One keyword, one target page | One question fanned out into sub-questions |
| Success signal | Position, clicks, impressions | Citation, mention, share of answers |
| Off-site role | Links | Links and mentions across sources the model reads |
| Feedback loop | Daily rank data | Manual, non-deterministic spot checks |
We mapped the conceptual version of this comparison in AEO vs SEO. This piece is the operating-model version: what changes on Monday morning.
What stays the same?
The technical foundation is unchanged, and it is still the majority of the work. Crawlability, indexation, site speed, clean information architecture, internal linking, and genuine subject authority all still decide whether you are in the candidate pool at all. AI systems overwhelmingly draw from content that conventional search already surfaces — if you are not indexed, you are not cited.
This is where most "SEO is dead" takes fall apart. Sites that were invisible to Googlebot are also invisible to GPTBot, PerplexityBot and Google-Extended. The failure mode is identical, and so is the fix. We hit this on our own site: a client-side React build that rendered fine for humans and returned an empty shell to crawlers — the fix and the schema patterns are in JSON-LD for React SPAs.
What changes in the actual work?
Four things change in the day-to-day: content structure, question coverage, off-site presence, and measurement.
Content structure. Traditional SEO tolerated a warm-up paragraph before the answer. AI SEO does not. The first sentence under every heading has to work as a standalone, quotable answer, because that sentence is the unit a model lifts. Headings phrased as buyer questions, one idea per section, short paragraphs — this is formatting as a ranking factor, in the literal sense.
Question coverage over keyword targeting. AI systems decompose a question into sub-questions, search each, then synthesise. A page that answers only the head term gives the model one thing to cite. A page that also covers the timeline, the cost drivers, the comparison and the risks gives it five. Cluster architecture — a pillar plus supporting posts that interlink — does the same job at site level; the AEO maturity model is a useful way to see where you currently sit.
Off-site presence carries more weight. Classic SEO cared about links as a ranking signal. AI SEO cares about mentions as a source signal: review sites, category roundups, Reddit and LinkedIn threads, comparison articles. If the "best X for Y" lists in your category do not name you, an AI answer about your category can be built without you — regardless of how strong your own site is.
Measurement is rebuilt from scratch. There is no rank tracker for a conversation. Answers vary by phrasing, by session, by model version. The workable substitute is a monthly manual log: take your ten most important buyer queries, run them in Google, ChatGPT and Perplexity, and record whether an AI answer appeared, whether you were cited, and who was cited instead. Presence or absence is the metric. It is cruder than rank tracking and considerably more honest.
Should I stop doing traditional SEO?
No — and framing it as a choice is the expensive mistake. Traditional SEO determines whether you are eligible; AI SEO determines whether you are chosen. Dropping the first to chase the second removes you from the candidate pool that AI systems draw from.
The sequencing that works: fix the technical foundation first, restructure content for extractability second, build off-site presence third, and stand up the measurement log alongside all of it. Every one of those steps also improves conventional rankings, which is why the trade-off people brace for mostly does not exist. For the engine-specific tactics, see how to show up in ChatGPT and Perplexity and AI overview optimization for B2B.
Where to start
Start by finding out whether you are cited today, because the answer usually settles the argument faster than any deck. Run your five highest-intent buyer queries through ChatGPT, Perplexity and Google, and write down who gets named. Then check one thing on your own side: view-source on your most important page and confirm the content is in the HTML, not assembled by JavaScript after load.
If you would rather not do that manually, book a free AI-search audit — we run the queries, check what the crawlers actually receive, and send back the gaps in priority order.