What is AI search?

AI search is any search experience where a language model writes the answer, drawing on live web results, instead of only returning a ranked list of links.

Want to know if your site gets AI search right?

I’ll check how AI search sees your business today and what would help it cite you, free.

No newsletter. Just what you asked for, from me. Privacy

In plain English

Examples include ChatGPT search, Perplexity, Google’s AI Overviews and AI Mode, Microsoft Copilot, Gemini and Claude with web search. Many appear to follow a similar pattern: the model turns the question into one or more searches, reads some of the results, then writes an answer with citations.

That retrieval step matters. It usually runs on a search index, either the provider’s own or a partner’s (Copilot uses Bing, for example), so pages that are crawlable, fast and clearly written have a head start. The model’s training data shapes the tone and background, while live retrieval likely shapes many of the specific businesses it names.

Why it matters for a small business

People ask AI tools longer, more specific questions than they type into Google. The answer often names only a handful of businesses. For most local businesses, Google still sends far more visitors today, but the share coming through AI tools is growing and worth preparing for. I wrote more on this in what makes a website rank in 2026.

An example

Someone in Edmonton asks Perplexity which bookkeepers in the city work with Shopify stores. The answer cites a small firm whose services page says exactly that, in a sentence, with a short list of the platforms it supports. A competitor that does the same work but describes itself only as “full-service bookkeeping” isn’t mentioned. (More for bookkeepers.)

Common mistakes

Where I see AI search go wrong most often.

  1. Building the site so content only appears after JavaScript runs, when many AI crawlers appear not to run JavaScript.

  2. Checking one AI tool once and drawing conclusions, when answers change between sessions and users.

  3. Abandoning Google work in a rush toward AI, even though Google still drives most local traffic.

FAQ

AI search, answered

  1. Often from live web searches run at the time of the question, combined with what the model learned in training. The live part is what you can influence.

  2. Partly. It still depends on crawling and indexing, but instead of ranking ten links it reads several pages and writes one answer, naming fewer sources.

Keep reading

Let’s begin

Let’s build something you’re proud to share.

Let’s Talk