AI search versus traditional SEO: what actually changes for local businesses

If you already invest in SEO, you do not need to throw it away. But you do need to understand how AI search changes the rules, so you can decide where to spend your effort. This article covers the mechanics: how AI search actually works under the hood, and where it diverges from the Google ranking model you already know.

How traditional SEO rewards you

In traditional SEO, Google crawls the web, builds an index, and ranks pages. When someone searches “estate agent Marbella”, Google returns an ordered list of links. Your reward for good SEO is a higher position on that list. The closer to position one, the more clicks you get.

The unit of success is a click. You optimise pages, earn links, and climb the rankings. The customer does the rest by clicking through to your site.

How AI search changes the reward

In AI search, the engine still crawls and reads the web, but it does not return a list of links. It generates a prose answer. When someone asks “who is the best estate agent in Marbella for British buyers?”, ChatGPT writes a paragraph naming two or three agencies and explaining why.

The unit of success is no longer a click. It is a mention. The customer may never visit your website, because the answer gave them enough to act on. They might message the named agency on WhatsApp, or ask a follow-up question, or simply remember the name for later.

This is the fundamental shift. You are no longer optimising for a position in a list. You are optimising for inclusion in a sentence.

The retrieval layer: how AI picks its sources

This is the part most SEO guides skip, and it matters. When you ask ChatGPT a question, it does not read the whole internet in real time. It runs a retrieval step first. It searches the web, fetches a handful of pages it thinks are relevant, and reads only those. Then it writes its answer based on what it read.

That retrieval step behaves a bit like Google Search, but with important differences. It tends to favour pages that answer the question directly, in plain language, with specific facts. It favours authoritative third-party sources over self-promotional ones. And it is sensitive to freshness: for questions where recency matters, it prefers recent content.

What this means for you: the pages that rank well in Google are not always the pages AI retrieves. A forum thread with a direct answer can outrank your polished homepage. A review site can be retrieved ahead of your own service page. The retrieval layer has its own preferences, and they are not identical to Google’s ranking algorithm.

The synthesis layer: how AI writes the answer

Once the model has retrieved and read its sources, it writes the answer. It synthesises what it found into a few sentences. It tends to name businesses that multiple sources agree on, described in similar terms. It tends to hedge when sources conflict. It tends to favour businesses it can describe specifically: “the estate agent for Scandinavian buyers in Mijas” is easier to synthesise than “a leading real estate agency”.

This is why consistency across sources matters so much. If your business is described one way on your website, another way on Google, and a third way on a directory, the model has to reconcile conflicting information. It becomes less confident naming you. If every source describes you the same way, the model can state it with confidence.

Where SEO and AI search agree

The good news is that the foundations overlap. Both reward:

• A fast, crawlable website with clear, useful content.
• A complete and accurate Google Business Profile.
• Genuine reviews from real customers.
• Mentions and links from trusted, relevant sites.

If your SEO foundations are weak, fix them first. They help both your Google rankings and your AI visibility.

Where they diverge

They diverge in three places that matter:

Position versus mention. SEO rewards a high position in a list. AI search rewards a mention in a sentence. A page can rank well and never be named, or be named without ranking well.

Your site versus the web. SEO is mostly about your own website. AI search weighs third-party sources heavily, because the model wants corroboration. Earning mentions in directories, press and reviews matters more than it used to.

Rankings reports versus answer tracking. A monthly report of Google positions tells you nothing about your AI visibility. You have to track AI answers directly, question by question, across engines. It is a different measurement.

Where to spend your effort now

Keep doing the SEO basics well. They still help. But add three things.

First, track your AI answers. Pick ten buyer questions and check them monthly across ChatGPT, Gemini and Perplexity. Second, work on the sources the models read: directories, review sites, press and your Google Business Profile. Third, make sure every source describes your business the same way. Consistency is the ranking factor AI search cares about most.

If you want help with any of that, our free check is the easiest starting point. We test ten questions across three engines and show you exactly where you stand.