More buyers now ask an AI assistant instead of typing into a search box: "Which companies supply saffron in bulk?" or "What is a good accounting system for a small shop?" The answer usually names only a few businesses. This guide explains, as far as anyone outside those companies can know, how that shortlist is formed and what you can do about it.
Two ways an assistant knows about you
An AI assistant draws on two sources, and they behave differently.
- What it learned in training. The model was built from a very large collection of text gathered up to a certain date. If your business was written about often before then, the model may know of it. You cannot change this after the fact, and it updates only when a new model is released.
- What it looks up live. Many assistants now run a web search when asked a question, read the top results, and write an answer from them, often with links to the sources. This part reflects the web as it is today.
For most small and medium businesses, the second source is where the opportunity is. It also means the foundations of ordinary search still apply: if a search engine cannot find and understand your page, an assistant that relies on that search engine will not either.
What seems to help
The companies that build these systems do not publish their exact rules, so treat this as informed observation, not a formula.
Clear statements of fact on your own site
An assistant has to extract facts quickly. A page that says plainly what you sell, where you operate, who you serve and how to reach you is easier to use than one written only in slogans. Question-and-answer sections work well because they match the way people ask.
Being mentioned in places it already trusts
When an assistant answers "who are the suppliers of X", it often leans on pages that list or compare suppliers: industry directories, review sites, marketplaces, trade publications and news articles. A business that appears on several of those is more likely to be named than one that exists only on its own website.
Consistent details everywhere
If your company name, address and description differ from one site to another, a machine may not realise they describe the same business. Consistency lets the mentions add up.
Structured data
The same machine-readable facts that help search engines help here: organisation, products, prices, location and opening hours, stated in a standard format.
Letting the crawlers in
AI companies use their own crawlers, with their own names. Some website security settings block them by default. If you want to appear in AI answers, check that you are not turning them away.
What nobody can promise
What can be done honestly is to ask the questions your buyers ask, repeatedly, and record what comes back: were you mentioned, which sources were cited, which competitors were named in your place. Over time those samples show a pattern you can act on, such as a directory that keeps being cited where you are not listed.
A practical order of work
- Make sure search engines can read your site. Our guide on what search engines read covers this.
- Write a clear page for each product or service, with the questions buyers ask answered in plain sentences.
- Get listed where your industry is listed. See which directories matter.
- Ask the assistants your buyers' questions each month and note what changes.
How GrowInSearch approaches this
GrowInSearch records dated samples of AI answers for the questions you choose, with the mentions and cited sources kept alongside. It does not report a single "AI rank", because that number would be made up. The AI monitoring page describes which assistants are covered and how.