AI SEO: How to Get Your Site Found and Cited by AI Search

When someone asks ChatGPT for a WordPress developer near Columbus, or asks Google a question that brings up an AI Overview, the answer they get is written from a handful of websites. AI SEO is the work of making your site one of them: easy for AI search to find, read, understand, and cite.

This guide replaces a run of shorter “AI SEO strategies” posts we published in 2025. They covered the same ground too many times, and some of what they said has dated. This is the one version we keep current: what AI search actually reads, the steps we take on client sites, how we measure it, and what nobody can promise you.

What “AI SEO” means

The term gets used for two different things:

• Optimizing for AI search. Getting your pages found, understood, and cited by ChatGPT, Google AI Overviews and AI Mode, Perplexity, Claude, and Microsoft Copilot. You’ll also see this called AI search optimization, generative engine optimization (GEO), or answer engine optimization (AEO). They all mean roughly the same thing.
• Using AI tools to do SEO. Keyword research, content drafts, audits, and reporting done faster with AI help.

Most of this guide is about the first one, because that’s what has changed. We cover the second near the end.

How AI search picks its sources

Every AI search product works a little differently, but the pattern is the same: look pages up, read them, write an answer, attach a few sources. Here’s what each of the big ones draws on.

• Google AI Overviews and AI Mode are built from pages in Google’s own index. Google’s guidance is plain: there are no extra requirements to appear in them. A page has to be indexed and eligible to show a snippet in regular search to be used as a supporting link.
• ChatGPT searches the live web and links its sources. OpenAI’s OAI-SearchBot is the crawler that makes a site eligible for ChatGPT search answers. OpenAI says sites that block it won’t appear in those answers.
• Perplexity searches the web for every answer and cites the pages it used. Its search crawler is PerplexityBot.
• Claude can search the web and cite what it finds. Anthropic’s Claude-SearchBot crawls for search; blocking it can reduce how visible you are in Claude’s answers.
• Microsoft Copilot draws on Bing, so a site that Bing can crawl and understand matters here too.

Two things follow from that. First, regular SEO still does most of the work: if Google and Bing can’t crawl and index a page, AI answers built on them won’t use it either. Second, there are a few new crawlers to let in, and a few new habits worth building. That’s where the steps below come in.

Step 1: Let the right crawlers in

AI companies run separate bots for separate jobs. Some fetch pages for search answers. Some collect pages to train models. Some visit a page only because a user asked about it. You can treat them differently, and you should decide on purpose rather than by accident.

Company	Search answers	Model training
OpenAI	OAI-SearchBot	GPTBot
Anthropic	Claude-SearchBot	ClaudeBot
Perplexity	PerplexityBot	None listed
Google	Googlebot	Google-Extended (a robots.txt setting)
Microsoft	Bingbot	None listed

Googlebot covers Google Search, including AI Overviews and AI Mode, and Bingbot covers Bing and Copilot. OpenAI, Anthropic, and Perplexity also have agents that visit a page only when a user asks about it (ChatGPT-User, Claude-User, Perplexity-User). OpenAI and Perplexity say robots.txt may not apply to those, since a person started the request.

Our default for clients who want to show up in AI answers: allow the search crawlers, and make a separate, deliberate call on the training crawlers. Blocking Google-Extended doesn’t affect your Google Search rankings or AI Overviews; Google says it only covers use in some of its other AI systems. A robots.txt along these lines does both:

# AI search crawlers: allowed
User-agent: OAI-SearchBot
User-agent: Claude-SearchBot
User-agent: PerplexityBot
Allow: /

# Model training: opted out (optional, your call)
User-agent: GPTBot
User-agent: ClaudeBot
User-agent: Google-Extended
Disallow: /

Then check the layer above robots.txt. Firewalls, security plugins, and CDN bot settings can block AI crawlers before robots.txt is ever read, and nobody notices because the site looks fine in a browser. When we take on a site, we check the server logs or the CDN’s bot report to confirm the crawlers we want are actually getting through.

Step 2: Make sure the content is in the HTML

Many AI crawlers read the HTML your server sends and don’t run JavaScript. If your services, prices, or answers only appear after scripts load (a tab widget, a page builder that renders in the browser, content pulled in by an app), those crawlers may see an empty page.

The quick test: open the page, choose View Source (not Inspect), and search for a sentence from the middle of the page. If it isn’t in the source, a crawler that doesn’t run JavaScript won’t see it either. On WordPress, a theme that renders pages on the server passes this test without any extra work. Speed matters for the same reason it always has: a page that responds slowly or times out is a page that doesn’t get read.

Step 3: Tell machines exactly who you are

An AI answer about “a WordPress agency in Dublin, Ohio” has to connect your website to a real business in a real place. Structured data (schema.org markup in JSON-LD) does that without any guesswork. On most business sites we add:

• Organization or LocalBusiness on the home page: name, logo, address, phone, and sameAs links to your Google Business Profile, LinkedIn, and other official profiles.
• Service on each service page, pointing back to the organization.
• Article or BlogPosting on posts, with a real author.
• FAQPage where a page genuinely answers questions. Google stopped showing FAQ rich results for most sites in 2023, so we add it for clarity, not for a search feature.

The markup only helps if it matches what’s on the page and what’s said about you elsewhere. Use the same business name, address, and phone number on your site, your Google Business Profile, and directory listings. Inconsistent details are one of the most common problems we find, and one of the cheapest to fix.

Step 4: Write pages that answer questions

AI assistants quote sentences that answer the question directly. Pages that bury the answer under three paragraphs of introduction, or never give a straight answer at all, are harder to cite. What works, in our experience:

• Answer first. Put the direct answer in the first sentence or two under each heading, then explain.
• Use the customer’s words in headings. “How much does WordPress maintenance cost?” is a heading an assistant can match to a question. “Our Approach” isn’t.
• Be specific. Prices, ranges, timelines, locations, and named steps give an assistant something concrete to repeat. Only publish numbers you can stand behind.
• One strong page per topic. Ten thin pages on the same subject compete with each other. We learned this one ourselves; it’s why this guide replaced the posts it did.
• Keep it current. Update pages when prices, services, or facts change, and show the date you last updated them.

A good way to find the questions: look at Search Console queries that are phrased as questions, read your contact form submissions, and ask your sales or support team what people ask on the phone. Then ask ChatGPT, Perplexity, and Google those same questions and see who gets cited.

Step 5: Show who’s behind it

Google’s quality rater guidelines describe what it looks for as E-E-A-T: experience, expertise, authoritativeness, and trust. AI search draws on the same signals in practice, because it draws on the same index. For a small or mid-size business that means a real About page, real author names on posts, real photos where you can, reviews on your Google Business Profile, and content written from work you’ve actually done. A paragraph about a problem you solved for a client (with their permission) is worth more than a page of general advice anyone could write.

Step 6: Don’t skip local

If you serve customers in a specific area, local search and AI search overlap heavily. Questions like “best HVAC company near me” or “web designer in Columbus” pull on map results, reviews, and business listings. The basics:

• Claim and complete your Google Business Profile: categories, services, hours, photos, and a description that says plainly what you do and where.
• Keep your name, address, and phone identical across the major directories and Bing Places.
• Ask happy customers for reviews, and reply to the ones you get.
• Give each location or service area its own page, written for that place.

Step 7: Add an llms.txt file (and keep expectations modest)

llms.txt is a plain-text file at the root of your site (yoursite.com/llms.txt) that summarizes who you are and links to your most important pages, written for AI assistants rather than people. The format was proposed in September 2024 at llmstxt.org.

We’ll be straight about it: no major AI company has committed to reading llms.txt, and Google says you don’t need any AI-specific file to appear in AI Overviews. We add it anyway because it takes little time, does no harm, and forces a useful exercise: deciding which pages actually matter. We built LLMs Ledger, a free llms.txt checker, generator, and directory, to make that quick. Treat llms.txt as one small part of the work, not the main event.

Using AI tools for SEO work

The other meaning of AI SEO is using AI to do the work. We use it every day, and it genuinely saves time on:

• Research: grouping hundreds of keywords by intent, and drafting the list of questions a page should answer.
• First drafts: outlines, meta descriptions, alt text, and schema markup, all reviewed by a person before they go live.
• Audits: summarizing crawl reports and spotting patterns in Search Console data.

Where it goes wrong is volume. Google’s spam policies treat mass-produced pages made mainly to rank as scaled content abuse, whether a person or an AI wrote them. AI-assisted content is fine when it’s accurate, edited, and useful to the reader. Every claim still needs checking, because AI tools will confidently state numbers and sources that don’t exist.

How to measure AI search

Measurement is still rough, so we use three sources together:

1. AI referrals in GA4. Visits from AI assistants show up as referral traffic. We add a custom channel group so they’re reported together, with a rule that matches the sources chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com, and claude.ai. Links clicked in ChatGPT also often carry utm_source=chatgpt.com.
2. Search Console. Clicks and impressions from AI Overviews and AI Mode are counted in your normal Google Search totals, with no separate filter. Watch the totals and the queries, not a dedicated AI report.
3. A monthly manual check. We keep a short list of the questions customers actually ask, put them to ChatGPT, Perplexity, Claude, and Google each month, and note whether the client appears and who gets cited instead. Answers vary from one run to the next, so we look for trends, not single results.

What nobody can promise you

No one controls what an AI assistant says. Anyone who guarantees you’ll be cited by ChatGPT, or promises a spot in AI Overviews, is guessing. What we can do is make your site as easy as possible to find, read, and cite, and then watch the data. We work to start seeing results within two to three weeks, and we adjust every month from what the numbers show.

A checklist to start with

• robots.txt allows OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot, and Bingbot, and you’ve decided on purpose about the training crawlers.
• No firewall, security plugin, or CDN rule is blocking the crawlers you want.
• Your key content is in the page source, not loaded by JavaScript.
• Organization or LocalBusiness schema on the home page, Service schema on service pages, and the same name, address, and phone everywhere.
• Each important question has one clear page that answers it in the first sentence or two.
• Real authors, an About page, and reviews on your Google Business Profile.
• Your site is verified in Google Search Console and Bing Webmaster Tools.
• An llms.txt file, checked with LLMs Ledger.
• An AI channel in GA4 and a short list of questions you check every month.

Want us to look at your site?

This is the work we do in our AI search optimization service, as part of our SEO plans or an AI consulting retainer. If you’d like to know where your site stands first, our free website check covers speed, search health, and AI readiness: whether AI crawlers can reach your site and what they’re missing. It’s a short report, not a sales call.