Listen instead

When someone asks ChatGPT “What’s a good chiropractor near me?” or “Find me a Baptist church in Virginia” — does your website come up? For most small businesses and churches, the answer is no. Not because AI doesn’t want to recommend you. Because it can’t find you.

A previous post on this blog raised the question of whether your site is invisible to AI at all. This one answers what to do about it. The good news: the fix is mostly about structure, not spending more money. You don’t need a new website. You need your existing site to speak a language AI understands.

Google AI Overviews now appear in roughly half of all US searches. ChatGPT handles over 700 million queries every week. Perplexity is the research tool of choice for a growing slice of professionals. If your site isn’t built for this — if it can’t be read, parsed, and cited by these systems — you’re missing a channel that grows larger every month.

// from the glossary: AEO · GEO

AEO (Answer Engine Optimization) is structuring content so AI can pull direct answers. GEO (Generative Engine Optimization) is optimizing for AI-generated search results. Both are about making your site visible to AI — not just Google.

How AI Search Is Different

Google crawls and indexes pages, then ranks them. AI reads, summarizes, and cites — or doesn’t cite. The distinction matters more than it sounds.

When ChatGPT answers a question, it draws on two things: what it learned during training (with a cutoff date, like a textbook published last year) and, in some modes, real-time web search. When Perplexity answers, it searches the web fresh for every single query and cites what it finds. Google AI Overviews pull directly from Google’s live index. Each platform has its own citation logic.

What they share is this: if your content isn’t clear, structured, and direct, AI skips you. A human visitor can skim a cluttered page and find what they need. AI can’t. It looks for patterns — headings that match questions, answers near the top of sections, entities stated plainly. If those patterns aren’t there, the system moves on to a source that has them.

Think of it as the difference between being in a phone book (Google) and being recommended by a knowledgeable friend (AI). The phone book just needs to find you. The friend needs to understand you well enough to vouch for you. Structure is what earns the voucher.

What AI Actually Looks For

Five things consistently show up in the research on what gets sites cited by AI systems.

Clear structure. Headings that match the questions people actually ask. Not “Our Services” but “What chiropractic conditions do you treat?” Not “About Us” but “What kind of church is Lighthouse Baptist?” AI reads headings the same way a researcher scans a document — looking for the section that answers the specific question in front of it.

Direct answers in the first sentence. Every section should answer its heading question immediately. Don’t warm up to the answer. Don’t save the key detail for the third paragraph. AI extracts the first relevant sentence it finds — if the answer isn’t there, it looks elsewhere.

Schema markup. This is structured data — code added to your site that tells AI and search engines exactly what your entities are. A LocalBusiness schema says: this is a business, here’s what it does, here’s where it is. A FAQPage schema gives AI a ready-made list of questions and answers to pull from. Without schema, AI is guessing what your site is about. With it, you’ve told it directly. See the schema markup guide for the specifics.

Entity clarity. Who you are, what you do, where you are — stated plainly, not buried. “We are a Baptist church in Waterbury, Connecticut, serving families in the greater New Haven area” is an entity statement. “Welcome to our community” is not. AI needs to build a clear picture of your entity before it can cite you accurately.

Fresh, authoritative content. Research from 2026 shows that visible date signals (a published date in the page markup) improve citation rates by roughly 30% on platforms like Perplexity. Thin pages — three sentences and a contact form — rarely get cited. Pages that answer a question fully and recently tend to get cited repeatedly.

// takeaway

AI doesn’t browse your site like a person does. It scans for structure, entities, and direct answers. If those aren’t clear, you’re invisible.

The Church Test

Try it yourself. Open ChatGPT or Perplexity and ask: “What are some good Baptist churches in [your city]?” Watch what comes up. Then ask why those churches appear and yours doesn’t — or why yours does.

Churches that surface in AI answers tend to share a few traits. They have a Google Business Profile that’s verified and complete, with service times, address, and denomination clearly stated. Their site has a real “About” page that names the denomination, describes the congregation, and uses plain language about beliefs and community. Their service information is in actual HTML text — not an image of a schedule, not a PDF download, not a slide in a graphic. Plain text that any system can read.

The church with the beautiful homepage built entirely from full-width images and decorative overlays is invisible. The pastor sees a stunning website. ChatGPT sees nothing it can read. For church websites, structure is not a design constraint — it’s the whole game.

The Business Test

Run the same test for a healthcare practice or local business. Ask Perplexity: “Who are the best chiropractors in [your city]?” The practices that get cited have a few things in common: clear service descriptions (not just service names), FAQ content that addresses real patient questions, and schema markup that tells AI exactly what conditions they treat and where they’re located.

“We treat pain” is not a citable statement. “We provide chiropractic care for lower back pain, neck stiffness, and sports injuries in Richmond, Virginia” is. One of those sentences an AI can use to answer a specific question. The other it can’t. For healthcare and wellness practices, specificity is the difference between being cited and being skipped.

AEO vs. GEO — Same Shift, Two Names

You’ll see both terms used in the industry, sometimes interchangeably. Here’s the practical distinction. AEO — Answer Engine Optimization — focuses on being the source AI cites when someone asks a specific question. “What does a chiropractor do?” “What do Baptists believe?” Your content either answers those questions cleanly enough to be pulled, or it doesn’t. GEO — Generative Engine Optimization — is the broader discipline of appearing favorably in AI-generated results overall, not just for specific answer pulls.

Both require the same foundation: structured content, schema markup, and entity clarity. Optimizing for one gets you most of the way to the other. Think of AEO as the precision layer and GEO as the full strategy. You need both — and neither requires rebuilding your site from scratch.

What to Actually Do

Five practical steps, in order of impact.

1. Structure content as Q&A. Go through your main service pages and ask: what question does each section answer? Rewrite the heading to match that question. Put the direct answer in the first sentence of the section. This alone is the highest-leverage change most sites can make.

2. Add schema markup. At minimum: LocalBusiness (or Church, MedicalBusiness, etc.), FAQPage on any page with question-and-answer content, and BreadcrumbList on every page. These are the signals that let AI understand your entities instead of guessing at them. The schema markup guide walks through exactly how to implement each type.

3. State your entity clearly. Every page should answer three questions in plain text: who you are, what you do, and where you are. Not buried in the footer. Not hidden in an image. In readable HTML, near the top of the page, in plain language.

4. Consider an llms.txt file. This is a simple text file placed at the root of your site — think of it as a roadmap for AI, like a sitemap for search engines. It tells AI crawlers what your site contains, how to interpret it, and where to find the most important content. It’s not yet a universal standard, but Vercel, Cloudflare, and Netlify all recommend it, and adoption is growing fast. For most small business sites, creating one takes about 30 minutes.

5. Run a visibility scan. The Artifex Visibility Report checks your site’s AI readiness across 30+ elements — schema, entity clarity, structured content, technical health — and shows you what AI systems actually see when they hit your pages. It’s free and takes about 30 seconds.

// takeaway

SEO gets you into Google’s index. AEO gets you into AI’s answers. You need both — and the foundation for both is good structure and clear content.

// FURTHER READING

Your Website Is Invisible to ChatGPT infographic
Why AI assistants can't see your site, at a glance