AI assistants now answer a large share of the questions that used to become clicks. They still cite sources. This is the work of becoming one of those sources.
Businesses noticing fewer clicks on informational searches, and anyone who has asked ChatGPT for a recommendation in their own industry and not appeared.
The work of getting your business cited by AI answer systems such as Google AI Overviews, ChatGPT, Perplexity and Copilot, instead of only ranking in classic search results. It overlaps heavily with good conventional SEO and adds a few specific requirements around structured data, content structure and crawler access. It is also called GEO or AIO, and our AI search page goes deeper.
Mostly it is normal SEO done well, plus three additions: machine-readable structured data, content written so a direct answer can be extracted cleanly, and verification that AI crawlers can reach your server. Anyone selling it as an entirely separate discipline is overstating it.
There is no submission form and no way to buy placement. What these systems draw on is what the web says about you: consistent descriptions of what you do, presence on sources they trust, structured data on your own site, and content that answers the question being asked. It is closer to reputation work than to a technical trick.
This is worth checking instead of assuming. Several sites we have tested allow AI bots in robots.txt while their CDN or firewall returns a 403 to those same bots at the edge. The owners had no idea. Our free audit performs a live fetch as an AI crawler, which is the only way to find it.
A real decision with a genuine trade-off, and it depends on your business. Publishers whose product is the content itself often block them. Businesses that want to be recommended almost always should not. What you should avoid is blocking them by accident, which is the common case.
An emerging convention: a plain-text file at your domain root summarising what your site is and what it contains, for AI tools that look for one. Adoption is early and it is not yet a standard anyone is obliged to honour. It costs almost nothing to add. Ours is at whatisseo.co.uk/llms.txt.
They reduce clicks on some informational searches and leave commercial and local searches largely intact. They also create a new position worth holding, which is being the cited source. The work shifts rather than disappearing.
Imperfectly, and anyone claiming precision here is overselling. You can test prompts relevant to your business across the main assistants and record whether you appear, track referral traffic from AI tools in Analytics where it is identifiable, and monitor whether your structured data is valid. We report all three and are clear about their limits.
Google judges the output instead of the method. What has been hit hard is high-volume content with no original expertise behind it, which describes most bulk AI publishing. Used as a drafting aid with a knowledgeable person editing, it is fine.
Organization and LocalBusiness so systems know who you are, Service or Product for what you sell, FAQPage for direct question-answer pairs, and Article with a modification date on editorial content. Connecting them into one graph rather than leaving them as separate blocks makes a noticeable difference to how reliably they resolve.
The technical parts are cheap and overlap with work worth doing anyway, so the downside is small. The upside is that competitors in most sectors have not looked at it yet.
Run the free audit. The AI-visibility section covers crawler access, structured data coverage and llms.txt, and it runs against your live server.