Heading Structure for AEO: How to Organize Pages AI Can Read

Most businesses treat headings as visual design choices. Bigger text for important stuff, smaller text for less important stuff. That approach worked fine when humans were the only ones reading your website.

It does not work for AI.

AI models like ChatGPT, Perplexity, and Google's AI Overviews do not see your page the way a person does. They do not scan layouts, notice bold fonts, or follow your eye-catching design. They read HTML. They parse the DOM. And the single most important structural signal they use to understand what your page is about is your heading hierarchy.

In our chiropractic AI citation study, we analyzed hundreds of pages across multiple AI models. The results were unambiguous: heading structure was the strongest predictor of whether a page earned AI citations. Not word count. Not backlinks. Heading structure.

This guide breaks down exactly how to organize your headings so AI can read, understand, and recommend your content.

Why heading structure matters for AI

Search engines have used heading tags as ranking signals for years. But AI answer engines treat headings differently than traditional search algorithms. Traditional SEO used headings as one of dozens of ranking factors. AI models use headings as their primary method of understanding page organization.

Here is why this distinction matters. When an AI model receives a user query, it needs to find not just a relevant page, but the relevant section of that page. It needs to extract a specific, coherent answer. Headings are how it navigates to the right section and determines whether the content beneath a heading actually answers the question.

Our research found that heading structure is the number one predictor of AI citations. Pages with clear, logical heading hierarchies were cited at dramatically higher rates than pages with flat or disorganized structures, even when the underlying content quality was comparable.

The reason is mechanical. Unlike humans who scan visually, AI reads the DOM structure literally. It treats your H1 as the definitive statement of what the page covers. It treats each H2 as a major subtopic. It treats H3s as supporting details. If your heading hierarchy is a mess, the AI's understanding of your page is a mess, and it will look elsewhere for a clean answer.

How AI reads your headings

AI models build an internal representation of your page based on heading structure. Think of it as a mental map. Each heading level creates a branch in a tree, and the content beneath it becomes a leaf. The clearer the tree, the easier it is for AI to locate the right branch when it needs an answer.

Here is how each heading level functions in AI's reading of your page:

• —H1 is the main topic of the page. It tells AI what the entire page is about. There should only be one.
• —H2s are the major sections or subtopics. Each H2 signals a shift to a new aspect of the main topic.
• —H3s are details within each section. They break down the H2 topic into specific sub-points.

AI uses this hierarchy to build a table-of-contents-like understanding of your page. When a user asks a question, the model can quickly navigate to the right H2, scan the H3s beneath it, and extract the paragraph that best answers the query.

Consider the difference between these two structures:

Bad heading structure

• —H1: Welcome to Our Chiropractic Clinic
• —H3: About Us (skipped H2 entirely)
• —H2: Why Choose Us
• —H2: Our Services
• —H2: Contact Us Today!
• —H1: Schedule Your Appointment (second H1)

Good heading structure

• —H1: Chiropractic Care in Austin, TX
• —H2: Chiropractic Services We Offer
• —H3: Spinal Adjustments
• —H3: Sports Injury Rehabilitation
• —H3: Pediatric Chiropractic
• —H2: What to Expect at Your First Visit
• —H2: Our Chiropractors
• —H2: Insurance and Pricing
• —H2: Frequently Asked Questions

The second structure creates a clean, navigable tree. AI can immediately understand what this page covers and jump to the right section. The first structure is ambiguous, has a skipped level, two H1s, and headings that describe the business rather than the information a searcher needs.

The heading structure AI loves

Based on our research, the pages that earn the most AI citations share a consistent heading pattern. Here are the principles behind it.

One H1 per page that clearly states the topic.

Your H1 should be a direct, descriptive statement of what the page is about. Not a tagline. Not a call to action. A clear topic declaration. If someone read only your H1, they should know exactly what information this page contains.

H2s that cover distinct aspects of the topic.

Each H2 should introduce a new, non-overlapping subtopic. Think of your H2s as chapters in a book. They should progress logically, each one covering territory the others do not. Avoid H2s that repeat the same idea in different words.

H3s that drill into specifics.

H3s should break an H2 section into concrete sub-points. If your H2 is 'Services We Offer,' your H3s should be the individual services. If your H2 is 'Common Causes,' your H3s should be each cause.

Each section should be self-contained enough to excerpt.

AI does not cite entire pages. It cites specific sections. The content under each H2 or H3 should stand alone as a coherent answer. If AI pulls just that section, it should make sense without requiring the reader to scroll up for context.

Headings should be descriptive, not clever.

AI does not understand puns, wordplay, or cultural references in headings. A heading like 'Cracking the Code' tells AI nothing about the section content. A heading like 'How Spinal Adjustments Relieve Lower Back Pain' tells AI exactly what the section covers. Be direct. Be literal. Save the creativity for your marketing copy.

Common heading mistakes that hurt AI visibility

These are the heading errors we see most frequently on local business websites, and each one directly reduces your chances of earning AI citations.

Multiple H1s on a page.

When a page has two or more H1 tags, AI cannot determine the primary topic. It receives conflicting signals about what the page is about. This is surprisingly common, especially on sites built with visual website builders that let you set any text size without regard for the underlying HTML tag. Every page should have exactly one H1.

Skipping heading levels.

Going from H1 directly to H3, or from H2 to H4, breaks the logical tree that AI is trying to build. The model expects a hierarchical structure, and gaps in that hierarchy create confusion. Always descend one level at a time: H1 to H2, H2 to H3.

Using headings for styling instead of structure.

This is the most pervasive mistake. Designers and site builders frequently use heading tags because they want bigger or bolder text, not because the content represents a structural heading. If a piece of text is not introducing a new section or subtopic, it should not be a heading tag. Use CSS for styling. Reserve heading tags for actual section headings.

Vague headings that do not communicate content.

Headings like 'Learn More,' 'Our Approach,' or 'Why Us' do not tell AI what the section contains. AI needs headings that describe the information in the section, not headings that describe the section's purpose from a marketing perspective. Instead of 'Why Us,' write 'Why Patients Choose Our Austin Clinic.' Instead of 'Learn More,' write 'Spinal Decompression Therapy: What It Treats.'

Putting important information in paragraphs without heading context.

If you have a section of valuable, answer-worthy content buried inside a long stretch of unheaded text, AI may not find it. Content that answers specific questions should be placed under a heading that signals what that content addresses. Headings act as signposts that tell AI where to look.

How to audit your heading structure

You do not need expensive tools to check your heading structure. Here is a straightforward process you can follow for any page on your website.

First, run your page through our AI Visibility Checker. It will flag heading structure issues automatically, along with other AEO factors that affect your AI citation potential.

For a manual audit, right-click on your page in any browser and select 'View Page Source' or open Developer Tools. Search for h1, h2, h3, h4, h5, and h6 tags. Read through them in order. They should tell a coherent story about the page's content.

Ask yourself these questions as you review:

• —Is there exactly one H1, and does it clearly describe the page topic?
• —Do the heading levels descend logically without skipping levels?
• —Does every major topic or subtopic have its own heading?
• —Could someone read only the headings and understand what the page covers?
• —Does each heading make sense as a table of contents entry?
• —Are any headings used for visual styling rather than structural organization?

If your headings pass all of these checks, your structure is solid. If not, prioritize fixing the issues in order: one H1 first, then correct hierarchy, then descriptive heading text.

Heading structure by page type

Different page types serve different purposes, and their heading structures should reflect that. Here are templates for the three most common page types on local business websites.

Service pages

Service pages are the most important pages for AI citations because they answer the questions potential customers are asking AI models. Your H1 should be the service name, and your H2s should cover the aspects of the service that people ask about.

Recommended structure:

• —H1: [Service Name] in [City, State]
• —H2: What is [Service Name]?
• —H2: Who Benefits from [Service Name]
• —H2: How [Service Name] Works
• —H2: [Service Name] Cost and Insurance
• —H2: Frequently Asked Questions About [Service Name]
• —H3: [Specific FAQ question 1]
• —H3: [Specific FAQ question 2]

This structure maps directly to the questions AI models receive. When someone asks 'How much does spinal decompression cost in Austin,' the AI can navigate straight to your cost section and pull a clear answer.

Blog posts

Blog posts should use headings to walk the reader (and AI) through the topic in a logical progression. Your H1 is the post title, H2s divide the post into major sections, and H3s add specificity within each section.

Recommended structure:

• —H1: [Post Title]
• —H2: [First major section]
• —H3: [Sub-point within the section]
• —H3: [Another sub-point]
• —H2: [Second major section]
• —H2: [Third major section]
• —H2: Frequently Asked Questions

Each H2 section should be substantial enough to serve as a standalone answer. For a deeper look at writing AI-friendly blog content, see our guide on how to write content AI can cite.

Location pages

Location pages need to establish geographic relevance while still providing useful, structured information. The H1 should combine your service and location. H2s should cover topics that establish local authority.

Recommended structure:

• —H1: [Service] in [City, State]
• —H2: [Service] Options in [City]
• —H2: About Our [City] [Practice Type]
• —H2: Our [City] Team
• —H2: Patient Testimonials
• —H2: [City] Location and Hours
• —H2: Frequently Asked Questions

The key for location pages is ensuring your headings include geographic context naturally. AI models use heading text to determine geographic relevance when answering location-specific queries.

Putting it all together

Clean heading structure is one of the highest-leverage changes you can make to your website's AI visibility. It requires no new content, no new pages, and no technical expertise beyond basic HTML understanding. Yet very few businesses do it well.

That is exactly what makes it a competitive advantage. While your competitors are using headings as visual decoration, you can use them as a structural roadmap that AI models follow directly to your content.

To recap the key principles:

• —One H1 per page that clearly declares the topic
• —H2s for each major subtopic, descending logically from the H1
• —H3s for specific details within each H2 section
• —Descriptive, literal heading text (no puns, no vague marketing language)
• —Self-contained sections that AI can excerpt as standalone answers
• —No skipped levels, no styling-only headings, no duplicate H1s

Start by auditing your most important pages. Use our AI Visibility Checker to identify structural issues, then work through the fixes systematically. For a comprehensive view of all the on-page factors that affect AI citations, review the full AEO checklist.

Heading structure is one piece of a larger website structure strategy for AI recommendations. But it is the piece that matters most. Get your headings right, and you give AI a reason to read the rest of your page. Get them wrong, and AI will skip you entirely, no matter how good your content is.

Frequently Asked Questions

Does heading structure matter more than content quality?

Both matter, but structure determines whether AI can find and extract your quality content. Think of structure as the delivery mechanism. You could write the best answer on the internet, but if it is buried in a wall of unstructured text, AI will not locate it. Structure makes quality discoverable. They work together, not in competition.

How many H2 sections should a page have?

There is no magic number. Use as many as needed to cover the topic thoroughly. Most effective pages have four to eight H2 sections, each covering a distinct subtopic. The goal is completeness without redundancy. If two H2s cover overlapping territory, combine them. If a section is trying to do too much, split it.

Can I use H4 and H5 headings?

Yes, but most pages do not need to go beyond H3. If you find yourself using H4 or deeper, consider whether the page is trying to cover too many topics. Deeply nested heading structures can signal to AI that a page lacks focus. For most local business websites, H1 through H3 provides all the structural depth you need.

Should I put keywords in my headings?

Write headings that clearly describe the section content. If that naturally includes keywords, great. Do not force keywords where they do not fit. Clarity matters more than keyword density. AI models are sophisticated enough to understand topic relevance without exact keyword matching. A heading that accurately describes its section will outperform a keyword-stuffed heading that confuses both AI and readers.

How quickly do AI models notice heading changes?

AI models re-index content on varying schedules. Most changes are reflected within two to six weeks. Making heading structure improvements is a medium-term investment, not an overnight fix. That said, the improvements tend to be durable. Once AI models learn that your page is well-structured, they are more likely to return to it for future citations.