Online store visibility in AI search: what should you do in 2026?

Why it matters whether you appear in AI search

Shoppers today ask ChatGPT, Perplexity or Google AI Overviews where to buy. If your online store is not among the recommendations, you do not even know about it: there is no click and no report. The good news: visibility depends on a few verifiable conditions.

AI search engines do not choose from backlinks. Citability is helped most by brand mentions, concrete numbers and quotations in the text, and basic crawlability. These three points are the backbone of the preparation; below we go through them in order, and we also tell you what you can skip.

The basics: crawlability and robots.txt

Before you write content, check whether AI crawlers can reach your pages. In robots.txt, search bots are often blocked by accident. The difference between training bots and search bots matters: GPTBot and ClaudeBot collect data for training the model, whereas OAI-SearchBot, Claude-SearchBot and PerplexityBot request data for search results. If you block only the training bot, you do not drop out of search. If you also block the search bot, you drop out.

For Google AI Overviews there is no separate opt-out that keeps you in the normal results. The nosnippet and max-snippet attributes are partial tools.

Check in the server log whether AI crawlers have downloaded your key pages at all. If they have not, every further step is pointless. The providers publish their IP lists, so logging is more reliable than referrer-based measurement.

The ChatGPT mobile app strips the referrer, so most of the traffic from there appears as direct visits in web analytics. That is why an accurate picture needs server-side logging and the providers’ IP lists; the dashboard report alone is not enough for this.

Structured data: what it is worth and what it is not

According to Google’s official documentation, there is no special AI schema that needs to be added. The Ahrefs experiment measured no improvement from structured data in AI Overviews citations, and only an uncertain increase in ChatGPT and AI Mode.

Structured markup of product, price and stock data remains useful in classic search, and Bing states that schema helps the language models of Copilot. Add the markup for the sake of classic search, but do not expect an AI breakthrough from it. The Web AI audit checks this layer too: it flags the pages where markup is missing or faulty.

Product pages and FAQ: the content that gets cited

AI looks for precise answers to narrow, concrete questions. A clear, citable answer to questions such as “what is the delivery time for product X” or “which size fits Y” is worth more than a general keyword article.

According to the GEO research (KDD 2024), adding quotations and concrete statistics increased visibility the most. Keyword stuffing reduced visibility below the baseline. In practice: write for real customer questions, and put concrete, dated numbers and named sources in the text.

FAQ content is the most important here. An answer to real questions typed into a chat is the easiest form to cite. It is worth starting from your own customer service questions, because they already use the customer’s words.

Keep product content up to date and machine-readable: price, stock, specifications. This certainly matters in classic search, probably in AI, and your assistant works from it too.

llms.txt: you can skip it

In spring 2026, Ahrefs examined the server logs of more than one hundred thousand domains: the vast majority of llms.txt files were never requested by anyone, and where there was traffic, it came mostly from SEO audit tools. Google’s position: “purely speculative for now”. The official crawler documentation of OpenAI and Anthropic does not mention it.

What to do instead: review robots.txt, check the server log and put your content in order. llms.txt does no harm, but do not start with it.

Brand mentions: they move more than another link

Brand mentions on the web and on video platforms correlate most strongly with AI visibility, independently of backlinks. This is a correlation; a causal link has not been proven. The pattern, however, is consistent. In practice, whether your brand is talked about in reviews, trade articles and comparison lists moves more than additional links.

Strengthening the brand name is a measurable visibility factor today. A rise in CTR was also measured for branded searches, while informational searches showed a decline.

What the Web AI assistant does

The Web AI assistant checks your pages in an SEO and GEO/AEO audit: crawlability, robots.txt, structured data, content quality. The audit shows on which pages it is worth writing precise, citable answers, and where there is a technical obstacle.

It also helps with content creation: answers written for real customer questions, the specification data of product pages and the FAQ content all come from the knowledge base the company uploads. The assistant answers in the customer’s language, even if the knowledge base is in another language.

Steps in order

The order matters because without the technical basics the content stays invisible too. Crawlability is the condition, content is the value, brand mentions are the amplifier.

• Review robots.txt: search bots should not be blocked.
• Server log: have AI crawlers downloaded the key pages.
• Product page: price, stock, specifications up to date and machine-readable.
• FAQ and content: citable answers written for real questions.
• Brand mentions: reviews, trade articles, comparison lists.
• llms.txt: not needed; it does no harm either.

How long until it shows

There is no published Hungarian benchmark for this. What you can expect: within days the server log shows whether crawlers download the pages. The effect of building brand mentions can be measured in months. If someone promises a quick guarantee, it is worth asking them what they measure and how.

The Web AI audit follows this order: first it checks the technical basics, then the content, and finally it gives the recommended actions page by page. If you put the basics in order now, your content will be citable later.

Next step

The Web AI audit checks crawlability and content, with recommendations for each page. Request a demo on your own website, or see the prices. If you have a question, write to .

Methodology

This guide is based on public crawler documentation, GEO research results and our experience with our own Web AI audits. We did not carry over traffic figures circulating on the market without a source; we only describe what can be verified.

Changelog

• 2026-10-15: rewritten from the smartaichat.hu article “Does ChatGPT find you?”

Sources

• Ahrefs: AI Overviews citations and llms.txt usage (March and May 2026) · ahrefs.com
• Google Search Central: AI features and your website (10 December 2025) · developers.google.com
• Aggarwal et al.: GEO – Generative Engine Optimization (KDD 2024, arXiv 2311.09735) · arxiv.org
• Own platform data, 90 days of use, 2026-07-08 – 2026-10-06 · own data · smartaiassist.eu/data/after-hours-questions/

Found a mistake, or do you have more recent data? Write to

and we will review it.

Frequently asked questions

Online store visibility in AI search 2026