AI visibility vs SEO: what changes, what stays the same

The surface is different

A classic search result is a list. Ten or so links are ranked, the person scans titles and snippets, and chooses. Your page competes for a click, and the position of that page is a stable fact you can look up.

A generated answer is a paragraph, a list or a comparison written by a model, often with a few sources attached. The person may never click anything. Your brand competes to be one of the few names mentioned, and the wording around that name matters as much as the mention itself.

Two consequences follow. First, there is no single rank to report. The same question can produce different answers on different days. Second, influence is indirect. You do not tell the engine what to say; you shape the material it draws on and the reputation it has seen.

What stays the same

The basics carry over almost unchanged. Pages must be reachable by crawlers, load without requiring a login, and contain real, specific, accurate information. Google's own guidance for AI Overviews and AI Mode says there are no extra technical requirements beyond what already applies to Search: a page has to be indexed and eligible to appear with a snippet to be used as a supporting link.

Authority still matters. Engines lean on sources they judge reliable, and those judgements correlate with the signals SEO has long cared about: reputable mentions, consistent facts across the web, clear authorship, and pages that stay accurate over time.

Intent matching still matters. A page that answers a question completely, in plain language, near the top of the page, is easier for both a human and a model to use.

Related reading: GEO vs SEO vs AEO: the differences and what to do and From keyword lists to answer landscapes.

What changes

Measurement changes first. Rank tracking assumes one result list per query. AI visibility needs sampling: you run a fixed set of prompts repeatedly across engines and record how often your brand is named, in what position among the named brands, with what sentiment and which citations. The result is an estimate with a range, not a position.

The unit of competition changes. In SEO you compete page against page. In AI answers you compete as an entity: the engine decides which brands belong in the answer, then looks for evidence about them. A strong page about a brand the engine barely knows may help less than a modest mention in several trusted places.

Third-party sources carry more weight than many teams expect. Engines often cite review sites, forums, publications and aggregators rather than the brand's own site, especially for comparison questions. If those sources say little or something outdated about you, the answer will reflect that.

Finally, the output is shaped by sub-questions. Google has described AI Overviews and AI Mode as using query fan-out: breaking a question into several related searches and combining what they find. A page that answers a narrow sub-question well can be used even if it would never rank for the original query.

Rankings do not predict mentions

It is tempting to assume that ranking first for a keyword means being named in the answer to that keyword. In practice the link is loose. Engines may draw on different sources than the top ten results, and some, such as chat assistants with search features, use their own retrieval and summarisation steps.

The practical test is simple: take your twenty most important commercial queries, ask them as a buyer would across the engines your market uses, and note whether your brand is named. Many teams with strong organic rankings find gaps. Many with modest rankings find they appear more often than expected, because of how they are described elsewhere.

A division of labour that works

Keep SEO owning crawlability, site quality, links and technical health. Add an AI visibility workstream that owns the prompt library, the sampling schedule, and the analysis of what engines say and cite.

Share the content roadmap. SEO supplies the keyword and intent research; AI visibility supplies the questions engines actually answer and the gaps where competitors are named and you are not. A page that fills one of those gaps usually benefits both channels.

Share the technical checks. Robots rules, rendering and structured data affect both. One extra item is specific to AI: which AI crawlers you allow, which is covered in our guide to AI crawlers, robots.txt and llms.txt.

A realistic first month

Week one: build a prompt library of 30 to 60 questions across discovery, comparison and branded intents, in the markets you serve. Week two: run it across the engines that matter to your audience and store full answers and citations.

Week three: read the answers. Which competitors appear most? Which sources are cited? What claims are made about you, and are they right? Week four: pick three fixes with the clearest evidence behind them, for example a page that answers a recurring comparison question, a correction on a source that is wrong about you, or a crawler block that is hiding your best pages.

Then keep the prompt library fixed and repeat. The point is a trend you can defend, not a one-off score.