GEO best practices for a WordPress site

This guide is for people who run a WordPress site and want it to show up in AI answers from Google, ChatGPT, Perplexity and Copilot. It walks through the WordPress settings, plugins and hosting layers that decide whether AI search can read your pages. A first pass takes an afternoon if you have admin access and can see your CDN or firewall settings.

The short version

1. Make sure WordPress itself isn’t asking search engines to stay away.
2. Check the robots.txt that WordPress and your plugins actually serve, and allow the AI search crawlers.
3. Check that your security plugin, host and CDN don’t block those crawlers.
4. Keep one XML sitemap, from core or one SEO plugin, and submit it.
5. Let one plugin output structured data, and keep it in line with what the page shows.
6. Make sure themes and page builders put the main text in the HTML.
7. Measure with crawler logs and the search engines’ AI reports.

1. Turn off “Discourage search engines”

The most common WordPress block is one checkbox. Staging sites are set up with it on, and it often stays on after launch.

WordPress’s Settings › Reading screen has a “Discourage search engines from indexing this site” option. Since version 5.3, ticking it adds a noindex,nofollow robots meta tag to every page that calls wp_head(), which standard themes do. WordPress notes that the option doesn’t block access. It is a request that search engines may honour.

Officially documented · WordPress.org

To appear as a supporting link in AI Overviews or AI Mode, a page “must be indexed and eligible to be shown in Google Search with a snippet”.

Officially documented · Google Search Central

A noindex page fails that test. Check the box is clear, then view the source of a live page and search for noindex.

2. Read the robots.txt you actually serve

WordPress generates a virtual robots.txt when no file exists. SEO plugins can edit it, and a physical robots.txt file on the server replaces it. Open https://your-site/robots.txt and read what is there, not what a settings screen says.

OpenAI: “Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though can still appear as navigational links.” Disallowing GPTBot, its training crawler, is a separate setting and does not affect search.

Officially documented · OpenAI

• Allow Googlebot, Bingbot, OAI-SearchBot, Claude-SearchBot and PerplexityBot if you want to appear in those products.
• Decide training crawlers (GPTBot, ClaudeBot, Google-Extended) separately. Rules per token: Configure robots.txt for AI crawlers.
• Watch for “block AI bots” toggles in plugins. Check which tokens they actually write.

3. Check the security plugin, host and CDN

A robots.txt that allows a crawler doesn’t help if a firewall answers it with 403 or a challenge page. On WordPress, that firewall can be a security plugin, the host’s bot protection, or a CDN in front of the site.

Google lists “ensuring that crawling is allowed in robots.txt, and by any CDN or hosting infrastructure” among the practices that still matter for its AI features.

Officially documented · Google Search Central

Cloudflare’s legacy “Block AI bots” setting blocks verified bots it classifies as AI training crawlers, and is marked for deprecation on September 15, 2026. From that date, new domains get defaults that block training and agent bots on pages that show ads, while search crawlers stay allowed. Mixed-purpose crawlers that do both search and training are blocked under the AI-training options.

Officially documented · Cloudflare

• List every layer between the internet and WordPress: CDN, host firewall, security plugin, rate limiter.
• In each, find the bot or AI setting and note what it blocks.
• Confirm in logs that verified search crawlers get 200 responses (step 7).

On WordPress, the robots.txt is rarely the only gatekeeper. Check every layer that can say no before you write a word of new content.

4. One sitemap, submitted

Since WordPress 5.5, core provides a sitemap index at /wp-sitemap.xml. It lists public post types, taxonomies and author archives, and the virtual robots.txt points to it. Core sitemaps are switched off when the site discourages search engines. Each sitemap holds up to 2,000 URLs by default.

Officially documented · WordPress.org (Make WordPress Core)

SEO plugins usually replace the core sitemap with their own. Pick one. Then:

• Check the sitemap URL loads and lists your important pages, not tag archives or attachment pages.
• Submit it in Google Search Console and Bing Webmaster Tools. Bing feeds Copilot.
• Background: XML sitemaps and XML sitemaps vs IndexNow.

5. One structured data source, matching the page

SEO plugins output schema.org markup automatically. Problems start when a theme, a plugin and a page builder each add their own, or when the markup describes things the page doesn’t show.

Yoast SEO outputs a connected schema graph as JSON-LD automatically. The types it lists include Organization, Person, WebPage, Article and Breadcrumb.

Officially documented · Yoast

“Don’t mark up content that is not visible to readers of the page.” Google also does not guarantee that correct markup will appear in results.

Officially documented · Google Search Central

Google says there is no special schema.org structured data needed to appear in AI Overviews or AI Mode.

Officially documented · Google Search Central

So treat structured data as housekeeping. Use it to state who publishes the site (Organization structured data) and what each page is (Article structured data). Whether it changes AI citations is an open question: Does structured data change AI search citations?.

• Run a page through the Rich Results Test and count how many plugins emit markup.
• Turn off duplicate output in the theme or extra plugins.
• Fill in the organization name, logo and profiles in the SEO plugin’s settings.

6. Keep the text in the HTML

Classic WordPress themes render on the server, so the text is in the HTML. Some sliders, tabs, “load more” blocks and headless front ends load content with JavaScript.

Google recommends server-side or pre-rendering because “not all bots can run JavaScript”.

Officially documented · Google Search Central

Fetch a page with curl and search for a sentence from the middle of the article. If it’s missing, see JavaScript sites (Next.js, Vue and others): making sure AI crawlers can read your content.

Two more template checks:

nosnippet prevents content from being used as a direct input for AI Overviews and AI Mode, and max-snippet limits how much may be used.

Officially documented · Google Search Central

Some SEO plugins expose these as per-post or site-wide “advanced robots” options. Search the page source for nosnippet, max-snippet and data-nosnippet. See Snippet controls for AI Overviews and AI Mode.

A setting you forgot about can undo a year of writing. Audit the template once, then audit it again after every plugin update.

What not to do

• Don’t install several SEO plugins at once. They output competing sitemaps, canonicals and markup (see step 5).
• Don’t block every “AI bot” in one click. Search crawlers and training crawlers are separate decisions, and blocking OAI-SearchBot removes you from ChatGPT search (step 2).
• Don’t treat an llms.txt plugin as a fix. There is no evidence it improves AI search visibility: Claim: llms.txt improves AI search visibility.
• Don’t mass-produce posts with AI plugins. Volume made to manipulate rankings is a spam policy issue: Scaled content abuse.
• Don’t mark up FAQs, reviews or ratings that aren’t on the page (step 5).

How you’ll know it’s working

Google’s crawlers can be verified with a reverse DNS lookup followed by a forward lookup, or against Google’s published IP lists.

Officially documented · Google

• Crawler access. In server or CDN logs, count verified visits from Googlebot, Bingbot, OAI-SearchBot and PerplexityBot, and their status codes: Verify AI crawler traffic in server logs.
• Google. Use the Generative AI performance report in Search Console: Generative AI performance report: AI Overviews and AI Mode in Search Console.
• Bing and Copilot. Use Bing’s AI Performance report: Bing AI Performance report.
• Answers. Run a fixed set of prompts monthly and record citations: Measuring AI visibility by sampling prompts.

Bing Webmaster Tools has an AI Performance report showing how often a site is cited in Copilot and Bing’s AI-generated answers.

Officially documented · Microsoft Bing

The Generative AI performance report “shows data about how your site performs in generative AI features on Google Search”.

Officially documented · Google Search Console Help

Evidence at a glance

Recommendation	Evidence
Clear “Discourage search engines”	Official (WordPress, Google)
Allow AI search crawlers in robots.txt	Official (OpenAI)
Check CDN, host and security plugin	Official (Google, Cloudflare)
One submitted sitemap	Official (WordPress)
One structured data source that matches the page	Official (Google, Yoast)
Structured data raises AI citations	Unestablished
Main text in server HTML	Official (Google)
Measure via logs and AI reports	Official (Google, Bing)