What "AI search visibility" actually means
People now ask questions of tools like Google's AI Overviews and AI Mode, ChatGPT, Perplexity, Claude and Gemini, and get a written answer instead of a list of links. Those answers are built from web pages. "AI search visibility" is simply how likely your site is to be read, used and named in those answers.
There is a lot of noise about this, with new acronyms every few months: GEO, AEO, LLMO. For a business with a WordPress site, the practical questions are much simpler. Can these tools reach and read your pages? Do your pages answer real questions clearly? And do they say who you are and why you are credible?
What Google actually says
It helps to start with the official position from the biggest player. In May 2026, Google published a guide on optimizing a site for generative AI features in Search, as reported by Search Engine Journal. Its main message is that this work is "still SEO." Google's AI features rest on its normal ranking and quality systems, so the fundamentals still decide who shows up.
On the two topics people ask about most, the guide is direct. Google says you do not need to create llms.txt files, special AI text files, or special markup to appear in its generative AI features. And while Google still recommends structured data as part of ordinary SEO, mainly for rich results, it says structured data is not required for these features and there is no special schema to add for them.
It also asks for pages that are indexed and eligible to appear in snippets, and for content that adds something beyond common knowledge, which it calls "non-commodity" content. It says you do not need to chop your writing into tiny pieces for AI systems, and you do not need to chase every long-tail phrasing.
One caution: this is Google's guidance for Google's products. ChatGPT, Perplexity and others may weigh things differently, and they do not publish their rules. Treat Google's guide as the strongest evidence available, not the final word on every AI tool.
Start with the boring part: can AI find and read your site?
Before schema or llms.txt, check the basics, because a site that cannot be crawled cannot be cited. On WordPress, these are the usual problems:
- The "discourage search engines" setting is on. It is under Settings, Reading, and it is easy to leave switched on after a launch.
- A noindex tag. An SEO plugin setting or a copy of a staging site can tell search engines not to index pages you want found.
- Blocked crawlers. A firewall, security plugin, CDN bot rule or robots.txt can stop crawlers from reaching you, sometimes by default.
- Content hidden behind scripts. Text that only appears after heavy JavaScript runs can be missed.
- Slow or unreliable pages. A site that times out or goes down is a site that gets skipped.
If you have never checked these, do it before anything else. It is the most common reason a good site is invisible.
Schema markup on WordPress
Schema is a labeling system. Behind the scenes, it tells machines things your design shows to people: this is a business called X, this page is an article written by Y on this date, these are the questions and answers on the page. It is written in a format called JSON-LD, and it does not change how your page looks.
For a typical business site, the types worth having are:
- Organization or LocalBusiness with your name, logo, address and profile links.
- Article or BlogPosting with author and dates on blog posts.
- Service for the main things you sell.
- BreadcrumbList to show how pages relate.
- FAQPage for genuine question-and-answer sections. Google now shows FAQ rich results only for a narrow set of sites, so do not expect fancy snippets, but the questions themselves still make your content easier to read for people and machines.
Most WordPress sites get the basics from an SEO plugin like Yoast, Rank Math or SEOPress. The usual mistakes are duplicate markup from the theme and a plugin both outputting the same thing, markup that says something different from what is on the page, and outdated details like an old address. Check your pages with Google's Rich Results Test or the Schema Markup Validator, and keep schema honest. Markup that makes claims your page does not support helps nobody.
llms.txt: what it is and whether to bother
llms.txt is a plain text file at the root of your site, at yoursite.com/llms.txt, that gives AI tools a short, readable summary of who you are and links to your most important pages. It was proposed as a convention to make a site easier for language models to understand. It is not an official web standard, and as noted above, Google says it is unnecessary.
We publish one ourselves, and we want to be straightforward about why. It takes little effort, it forces us to state clearly what we do and which pages matter, and it does no harm. We do not claim it improves rankings, and nobody has shown strong evidence that the major AI tools rely on it. If you add one, keep it accurate, keep it short, update it when your pages change, and do not use it for keyword stuffing. It is worth an hour. It is not worth a strategy.
Structured content: the part that matters most
If schema labels your page for machines, structured content is how the page itself is built. This is where most of the real gains are, and it is also just good writing for people.
- Lead with the answer. Put a short, direct answer near the top, then explain. Readers and AI tools both find it faster.
- Use headings that match real questions. "How much does X cost?" is clearer than a clever title.
- Keep paragraphs short and specific. Concrete details beat general statements.
- Use tables and lists where they help, such as comparisons and step-by-step instructions.
- Add real expertise. First-hand experience, specific numbers you can stand behind and honest limits make a page worth citing. Generic content that repeats what every other site says is what Google calls commodity content.
- Show who is behind it. Name the author, explain their background, and keep your business details consistent everywhere.
- Keep it current. Update posts when facts change, and show the date.
Notice that none of this is specific to AI. That is the point of Google's "still SEO" message. A page that serves a human reader well is the best preparation for any tool that reads pages on a human's behalf.
A short checklist for a WordPress site
- Confirm search engines are allowed to index the site, and that staging copies are not indexed.
- Check that your firewall, security plugin and CDN are not blocking the crawlers you want.
- Submit an XML sitemap and make sure it lists the pages that matter.
- Install or review your SEO plugin's schema output and test your key pages.
- Rewrite your most important pages to lead with a clear answer.
- Add an author and a short bio to blog posts.
- Optionally publish an llms.txt, and keep it accurate.
- Keep the site fast, secure and online, because none of the above matters if it is down.
How to tell whether it is working
Measurement is the weak spot, and anyone who promises exact AI visibility numbers is overselling. Google Search Console does not separate AI Overviews from normal search results. What you can do is watch your analytics for referral traffic from AI tools, such as ChatGPT and Perplexity, ask those tools the questions your customers ask and see whether your business is named, and keep an eye on branded search over time. It is rough, but the trend over a few months tells you more than any single check.
Where managed operations fits
Almost everything above depends on a technical foundation: pages that load fast, stay online, are not accidentally blocked, and carry accurate markup. That is ongoing work, and it is the kind that falls between IT, marketing and whoever built the site. Managed WordPress operations keeps that foundation in shape, with managed WordPress plans built around updates, security, monitoring and uptime.
If your issue is getting clear answers about how your site performs in search, our AI Automation Service turns your Search Console and Analytics data into a plain-English monthly report.
Sources
Google's position is reported by Search Engine Journal in its coverage of Google's May 2026 guide to generative AI features in Search.
