The Complete AEO Guide: What We Know, What We Don’t, and What to Do Now

A Practical Guide to AI Search Optimization From Someone Who’s Building It in Public

If you run a business that depends on people finding you online, something fundamental has changed. Google, ChatGPT, Perplexity, and a growing list of AI systems are answering your customers’ questions directly. No click required. No visit to your website. Just an answer pulled from whatever source the AI decided to trust.

That’s what Answer Engine Optimization is about: making sure your business is the source those systems trust. Not through tricks or hacks, but through content structure, credibility signals, and technical fundamentals that most businesses haven’t even heard of yet.

This guide covers what AEO actually is, what the research says works, what nobody can prove yet (even though plenty of people are selling it like they can), and exactly what you can do about it today. Written by someone who spent 32 years in plumbing operations before getting into digital marketing, which means you’ll get the truth, not a pitch.

By Tim Dini | Last updated February 2026

What Changed (and When Most People Weren’t Looking)

For twenty years, the game was simple. Someone types a question into Google, Google shows a list of websites, the person clicks one. Your job was to be as high on that list as possible. That was SEO. And it worked.

Here’s what changed: people stopped clicking the list.

It happened gradually, then all at once. Google started showing AI-generated summaries at the top of search results (AI Overviews). ChatGPT went from zero to 800 million weekly active users in about two years, making it one of the fastest-adopted technologies in human history. Perplexity built an entire search engine around AI-generated answers and is now processing hundreds of millions of queries a month. Voice assistants got smart enough to give real answers instead of “here’s what I found on the web.”

The result is that a growing percentage of searches now end without anyone clicking on a website. Similarweb’s data shows zero-click searches jumped from 56% in 2024 to 69% by mid-2025. The AI reads the web, synthesizes an answer, and delivers it directly. The user gets what they need. The websites that weren’t cited get nothing.

For businesses that depend on internet traffic for customers, this isn’t a minor shift. It’s a fundamental change in how people find you.

Seer Interactive tracked over 3,000 search queries across 42 organizations through 2025 and found that organic click-through rates dropped 61% on queries where Google showed an AI Overview. Major publishers like Forbes, HuffPost, and Business Insider reported traffic losses between 50% and 55%. Pew Research tracked nearly 69,000 real Google searches and found that only 8% of users who saw an AI Overview bothered clicking a traditional search result.

I watched one site’s organic traffic drop 79% over the course of 2025. Quarter over quarter, it just kept falling. Conversions followed, down 64% over the same period. Now, that’s an extreme case and additional factors like Google updates didn’t help, but it’s not a complete outlier.

I’d love to tell you exactly how much of that was AI search eating their clicks. I can’t. The tracking tools aren’t there yet, and anyone who tells you they can precisely measure AI search impact right now is either lying or selling something. Their rankings dropped during the same period too, so I can’t isolate one cause.

But here’s a question I keep coming back to: what does “ranking” even mean anymore? If an AI Overview answers the question at the top of the page and your #1 organic result is sitting below the fold where nobody scrolls, are you really ranked #1? You’re first in a list nobody’s looking at.

That’s what makes this moment so uncomfortable for businesses that depend on search traffic. The old metrics don’t tell the story they used to. Traffic is down, conversions are down, and the tools we’ve relied on for twenty years to explain why are only now starting to catch up with what’s actually happening.

This is exactly why I started paying attention to AEO. Not because I had it figured out, but because the numbers stopped making sense and I needed to understand why.

Why This Matters If You Run a Business

Let me make this specific. Say you’re a personal injury lawyer in Chicago. Someone asks ChatGPT “what should I do after a car accident in Chicago?” The AI gives a detailed answer and mentions two or three law firms by name as sources.

If your firm is one of those sources, you just got a referral from the most trusted tool your potential client uses. If you’re not, you don’t exist in that conversation. And here’s the painful part: it doesn’t matter how good your Google rankings are if the AI never pulls from your site.

There’s data to back this up. Seer Interactive found that brands cited within Google’s AI Overviews earned 35% more organic clicks and 91% more paid clicks than brands that weren’t cited. One AI citation can generate more qualified traffic than ranking #3 in traditional results. That’s not theory. That’s from an analysis of 25 million organic impressions.

This same scenario is playing out in every industry. Home services, insurance, healthcare, financial services, e-commerce. Anywhere a potential customer might ask an AI a question before making a decision, AEO determines whether your business is part of the answer.

The businesses that figure this out now have a window of advantage. Most of your competitors haven’t started. The ones who have are mostly doing it wrong because they’re following advice from people who’ve never tested it. That window won’t stay open forever, but right now, it’s wide open.

AEO vs. Traditional SEO: What’s Different and What’s the Same

Let me be clear about something: AEO does not replace SEO. If someone tells you to abandon your SEO strategy and go all-in on AEO, they’re either selling you something or they don’t understand how this works. Traditional search isn’t dead. It’s evolving. AEO is the evolution, not the replacement.

Google’s John Mueller said it plainly at Google Search Live in December 2025: “AI systems rely on search, and there is no such thing as GEO or AEO without doing SEO fundamentals.” I couldn’t agree more. Every guide that tells you AEO is some revolutionary break from SEO is misleading you. The revolution is in how the results are delivered. The foundation hasn’t changed.

Here’s what stays the same: you still need a technically sound website, quality content, genuine authority in your space, and backlinks from credible sources. Google’s core ranking factors haven’t disappeared. They still matter. In fact, an analysis of 432,000 keywords by seoClarity found that 97% of AI Overview citations come from pages already in the top 20 organic results. Traditional SEO ranking is still the primary gateway to getting cited by AI.

Here’s what’s different:

Content structure matters more than ever. AI systems need clear, well-organized content to cite accurately. A page that rambles for 2,000 words before getting to the point won’t get cited, even if it ranks well in traditional search. NVIDIA’s research on how retrieval-augmented generation (RAG) systems work shows that content chunked into standalone 200-500 word sections that each answer a specific question gets retrieved most accurately.

Schema markup matters, but not the way most people think. Structured data helps Google understand your content, which improves your traditional rankings, which in turn improves your chances of being cited in Google’s AI Overviews. That chain matters. But here’s something most AEO guides won’t tell you: a December 2025 study from Search Atlas analyzed schema coverage against LLM citation rates across ChatGPT, Gemini, and Perplexity and found that schema markup alone does not influence how often those platforms cite your content. Domains with extensive schema were cited no more frequently than domains with little or none.

This doesn’t mean schema is useless. It means you need to understand why it helps. Schema helps you rank better in Google. Google’s AI Overviews overwhelmingly cite pages from the top 10 organic results. So schema helps you get cited by Google’s AI, but through the side door of traditional ranking, not because ChatGPT is reading your JSON-LD.

For standalone AI platforms like ChatGPT and Perplexity, what matters more is your content quality, your brand authority, and whether your information shows up consistently across multiple credible sources on the web. I’ll get into all of that.

E-E-A-T signals carry more weight. E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness. AI systems are trained to prioritize authoritative sources. They’re looking for real expertise, real experience, and real credentials. A major AI visibility study analyzing over 680 million citations found that brand search volume, not backlinks, is the strongest predictor of whether an AI system cites you. The sites that have been faking authority with keyword stuffing and generic content are about to have a very bad year. If your business is in a YMYL industry (legal, medical, financial, home services), the standard is even higher. The YMYL guide on this site explains why.

Direct answers beat comprehensive pages. In traditional SEO, longer content often ranked better. In AEO, the page that answers the question most clearly and concisely in the first 100 words often wins the citation, even if a competitor has a longer, more comprehensive page.

The Big Three Work Differently (And That Matters)

This is something most AEO guides skip entirely, and it’s a mistake. Not all AI systems find and cite content the same way.

Google AI Overviews pull citations almost exclusively from pages that already rank well in traditional organic search. seoClarity’s data shows 97% of AI Overview sources come from the top 20 results. This means your traditional SEO work is your primary path to Google AI citation. Schema markup, strong backlinks, technical SEO, all of it feeds this channel.

ChatGPT works two different ways depending on whether web browsing is enabled. Without it, ChatGPT draws entirely from its training data, which means it’s pulling from what it learned during training. OpenAI’s training data prioritizes Wikipedia, licensed publisher partners, and then sites like Reddit and industry publications. With web browsing on, it queries Bing and selects 3-10 sources. Either way, what drives ChatGPT citations is brand authority. How often your brand gets mentioned across the web, whether you’re on Wikipedia, whether people discuss you on Reddit and in industry publications.

Perplexity has its own web index and emphasizes real-time content. It leans heavily on Reddit, news sources, and recently published content. Recency matters more here than on the other two platforms.

The practical takeaway: if you only optimize for Google, you’ll miss ChatGPT and Perplexity. If you only focus on content quality, you might miss Google’s AI Overviews which still care about traditional ranking signals. The businesses that get cited everywhere are the ones that have both strong traditional SEO and strong brand authority across multiple platforms.

“So What Do I Actually Need to Do?”

If you’re a business owner or manager reading this, here’s the honest summary of what AEO requires:

1. Restructure your existing content so answers come first. Take your best-performing pages and make sure the most important information is in the first few paragraphs, not buried halfway down. AI systems cite from the top of the page.

2. Add schema markup to your key pages. This is the technical part that most business owners will need help with. It’s not difficult, but it needs to be done correctly. Remember: schema primarily helps you through Google’s ecosystem, not directly with ChatGPT or Perplexity. But Google is still where most search happens, so it matters. See the Schema Markup guide on this site for the full walkthrough, or ask your developer or SEO team to handle it.

3. Build genuine authority signals, everywhere. Author bios, credentials, original data, expert quotes, real testimonials. AI systems are getting better at distinguishing real expertise from content-farm filler. But here’s the part most guides miss: your authority needs to show up across the web, not just on your own site. LinkedIn, Reddit, industry directories, YouTube, and third-party publications all feed the systems that AI uses to decide who’s credible. See the E-E-A-T guide on this site.

4. Create FAQ content based on real questions your customers ask. Structured FAQ pages with proper schema markup are one of the highest-impact AEO tactics. They directly match how people query AI systems.

5. Make sure AI can actually reach your site. Check your robots.txt and CDN settings. If you’re blocking GPTBot, ClaudeBot, or PerplexityBot, fix that before you do anything else.

6. Monitor your AI visibility. Use the new AI monitoring tools mentioned in this guide to check whether your business appears in AI-generated answers for your key queries. Note what comes up and what doesn’t. This is your new baseline.

That’s the practical overview. If you’re the person responsible for implementing this (or you want to understand the details before briefing your team), Part 2 below walks through every step.

Step 1: Audit Your Existing Content for AI Readiness

Before you build anything new, figure out what you’ve already got. Most businesses are sitting on content that’s 60-70% of the way to being AEO-optimized. It just needs to be restructured.

Pull up your top 20 pages by traffic. These are the pages you’re going to optimize first because they already have authority signals (backlinks, traffic, age) that AI systems value. Starting with new content is a mistake most people make. Start with what’s already working.

For each page, answer these questions:

Does the page answer a specific question in the first 100 words? If the answer is buried in paragraph six, the AI will probably cite someone else who put it up top.

Is the content structured with clear headings? AI systems use heading hierarchy (H1, H2, H3) to understand content structure. If your page is one long wall of text, you’re making the AI guess what each section is about. Use semantic HTML5 elements too (article, section, header, footer, time tags). These help AI systems parse your content structure beyond just headings.

Does the page have schema markup? Check using Google’s Rich Results Test (search.google.com/test/rich-results). If there’s no structured data, that’s your first fix.

Is there a clear author with credentials? Pages with identified, credentialed authors get cited more than anonymous content. This isn’t theory. Research shows that content with clear author transparency and unique firsthand experience performs significantly better in LLM citation rates. This pattern holds regardless of industry.

Is the information current? AI systems favor recently updated content. If your best page hasn’t been touched in two years, update it with current data and change the “last modified” date (honestly, by actually modifying it). And make sure your [dateModified] schema property reflects the real update date. RAG systems (retrieval-augmented generation, the technology that powers how AI platforms like ChatGPT and Perplexity pull information from the web) actively look for this signal.

Is your site accessible to AI crawlers? This should be step zero, honestly. Check your robots.txt for any rules blocking GPTBot, ClaudeBot, PerplexityBot, or Google-Extended. If you’re using Cloudflare or a similar CDN, check whether AI bot protection is turned on. Ahrefs studied roughly 140 million websites and found GPTBot is the most blocked crawler at about 6%. If the AI can’t read your site, nothing else on this checklist matters.

I’ll be honest with you: a clean, reliable AEO audit process is still being figured out. Traditional SEO audits have twenty years of tools and frameworks behind them. AEO audits are maybe six months into that journey. But the tools are finally starting to catch up.

Semrush now offers a full AI Visibility Toolkit (launched October 2025) that tracks which prompts generate citations for your site and lets you compare your AI visibility against competitors across ChatGPT, Google AI Overviews, AI Mode, Perplexity, and Gemini.

These aren’t perfect yet, but they’re real data from tools the industry already trusts. I’ll be actively testing these and building my own audit workflow around them. When I’ve used them long enough to know what actually correlates with results, I’ll publish a full review.

For now, combine the manual checklist above with whatever data these tools give you. It’s not elegant, but it’s a hell of a lot better than where we were six months ago.

Step 2: Restructure Content for AI Citation

This is where most of the impact comes from, and it’s not technically complicated. It’s an editing exercise.

The Answer-First Pattern

Every page on your site that targets a question should follow this structure:

First 100 words: Direct, clear answer to the primary question. No preamble, no “in today’s digital landscape” throat-clearing. Just the answer.

Next 200-300 words: Context and supporting detail. Why the answer is what it is. Evidence, data, or experience that backs it up.

Rest of the page: Comprehensive coverage for people (and search engines) that want the full picture. Related subtopics, implementation details, examples.

Here’s what this looks like in practice. Say you’re a roofer and you have a page about “How much does a roof replacement cost?” The old SEO approach was to write 2,000 words about roofing materials, factors that affect pricing, regional differences, and then maybe mention a price range in the conclusion. The AEO approach: start with the price range in the first sentence, then explain the factors.

Bad (traditional SEO approach): “When considering a roof replacement, there are many factors that homeowners should take into account. The type of roofing material, the size of your home, the complexity of the roof line, and regional labor costs all play a role in determining…” (answer buried 800 words in)

Good (AEO approach): “A full roof replacement typically costs between $8,000 and $25,000 for an average-sized home, depending on materials and location. Asphalt shingles fall on the lower end, metal and tile on the higher end. Here’s what drives the price for your specific situation…” (answer in the first 30 words)

That restructuring alone, putting the answer first, is the single highest-impact AEO change you can make to existing content.

One thing to keep in mind: NVIDIA’s benchmarks on RAG (retrieval-augmented generation) system performance show that individual paragraphs of 200-500 words that can stand alone as complete, citable units perform best for AI retrieval. Think of each major section of your page as a self-contained answer that an AI could pull and quote without needing the rest of the page for context.

Step 3: Implement Schema Markup

Schema markup is structured data you add to your website’s code that tells AI systems and search engines exactly what your content is about. Think of it as labeling boxes before a move. You could throw everything in unmarked boxes and hope the movers figure it out. Or you could label each box “kitchen,” “bedroom,” “fragile” and make everyone’s job easier.

AI systems process billions of web pages. Schema markup is how your page says “Hey, I’m a FAQ about personal injury law, written by a licensed attorney with 20 years of experience, last updated three weeks ago.” Without it, the AI has to guess all of that from context.

Now, I want to be upfront about something most AEO guides won’t tell you. Schema markup primarily helps you through Google’s ecosystem. It improves your chances of getting rich results, featured snippets, and inclusion in AI Overviews because it helps Google better understand your content, which improves your rankings, which makes you eligible for AI citation. That chain is real and it works.

But if you’re expecting schema markup to directly get you cited by ChatGPT or Perplexity, that’s not how those systems work. They use RAG to pull from search indexes and web content. They’re reading your text, not your JSON-LD. The Search Atlas study I mentioned earlier confirmed this across hundreds of domains.

Does that mean schema doesn’t matter? Absolutely not. Google is still where the vast majority of search happens. Google AI Overviews overwhelmingly cite top-ranking pages. Schema helps you rank better. Better ranking leads to AI citation. The connection is just less direct than most people claim.

The Schema Types That Matter Most for AEO

FAQPage: If you create one type of schema today, make it this one. FAQ schema directly matches how people query AI systems (as questions) and is one of the easiest to implement. Take the five most common questions your customers ask and create a properly marked-up FAQ section on your relevant pages.

Article / BlogPosting: Tells AI systems this is a substantive piece of content, who wrote it, when it was published, and when it was last updated. Every blog post and pillar page should have this. Pay special attention to the [dateModified] property. RAG systems actively look for freshness signals.

HowTo: For any step-by-step content. AI systems love this because it maps directly to procedural queries (“how do I…” questions).

LocalBusiness: For any business with a physical location. This is critical for local AEO. It tells AI systems what you do, where you’re located, your hours, your service area, and your contact info.

Person: For author bios and expert pages. This connects your content to a real human with real credentials. AI systems use this for E-E-A-T assessment.

Organization: Establishes your business entity and connects it to your content, your people, and your external profiles.

For detailed implementation instructions with code examples, see the Schema Markup for AI guide on this site. That guide walks through each schema type with actual JSON-LD code you can copy and customize.

Step 4: Build Genuine Authority Signals (Across the Entire Web)

Here’s the thing about E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) in the context of AI search: you can’t fake it. Not anymore.

In traditional SEO, you could “game” authority signals. Buy some backlinks, stuff some keywords, publish a lot of thin content, and sometimes rank for things you had no business ranking for (shameful). AI systems are different. They’re trained on massive datasets and they’re increasingly good at distinguishing between genuine expertise and content that’s just optimized to look like expertise.

But here’s what took me a while to fully understand: building authority for AI isn’t just about what’s on your website. It’s about what the entire web says about you. AI systems don’t just read your site. They cross-reference it against everything else.

HubSpot calls this “building consensus,” and I think that’s the right way to frame it. When multiple independent, credible sources say the same thing about your brand, your expertise, or your position on a topic, AI systems treat that information as more reliable and more citable. When only your own website makes a claim about how great you are, the AI is less likely to repeat it.

This is why some brands with mediocre websites but strong reputations across the web get cited more than brands with beautifully optimized sites that nobody else talks about.

Here’s what genuine authority signals look like in practice:

Named, credentialed authors on every piece of content. Not “Staff Writer.” Not “Admin.” A real person with a real bio, real credentials, and links to their profiles on other platforms. AI systems cross-reference this.

Original data and research. If you have unique data from your business (anonymized client results, industry surveys, original analysis), publish it. AI systems prioritize primary sources over sites that are simply summarizing other people’s data. One study found content featuring original statistics sees 30-40% higher visibility in LLM responses. If you’ve got unique numbers, use them.

Consistent presence across the web. Your author and business should appear on LinkedIn, relevant industry directories, YouTube, and platforms that feed AI training data like Reddit and Medium. This isn’t just a nice-to-have anymore. It directly affects whether AI systems consider you credible enough to cite.

Reddit deserves special attention here. Perplexity indexes Reddit heavily. ChatGPT’s training data includes Reddit content with meaningful engagement. Google’s AI Overviews frequently pull from Reddit discussions. If your brand or your experts are active in relevant subreddits with genuinely helpful contributions (not spam), that visibility feeds directly into AI citation.

YouTube is another one. Google AI Overviews frequently include video carousels, and YouTube content gets preferential treatment in those results. If you’re creating video content that answers the same questions your written content addresses, you’re doubling your surface area for AI citation.

Real testimonials and case studies. Not “Sarah Chevelle, VP of Marketing at TechFlow SaaS” (who doesn’t exist). Real names, real companies, real results. If you can’t share client names, anonymize them honestly: “A personal injury firm in Chicago” is more credible than a fabricated testimonial, because it’s true.

Third-party mentions and press. Getting quoted in industry publications, being mentioned in roundup articles, earning genuine press coverage. These aren’t just good for backlinks anymore. They’re brand signals that AI systems use to determine authority. A major 2025 analysis found that brand mentions are roughly 3x more important than backlinks for AI search visibility.

The full breakdown of E-E-A-T strategy for AI search is in the E-E-A-T guide on this site. But the key principle is this: AI systems are getting better at finding the real experts in any space. Be a real expert, make it easy for them to verify that you are one, and make sure the rest of the web confirms it.

Step 5: Build Strategic FAQ Content

FAQ content is the lowest-hanging fruit in AEO, and most businesses are leaving it on the tree.

Here’s why it works so well: when someone asks an AI a question, the AI is literally looking for a question-and-answer pair to cite. A properly structured FAQ page with FAQPage schema is exactly that. You’re handing the AI exactly what it needs, in exactly the format it wants.

How to Build FAQ Content That Gets Cited

Start with real questions. Not questions you made up because they contain your target keywords. Actual questions that real customers ask you. Check your email inbox, your call logs, your chat transcripts, your Google Business Profile Q&A. Those are your FAQ sources.

Answer each question in 40-80 words. Clear, direct, no filler. AI systems prefer concise answers they can cite in full. If your answer is 300 words, it’s not an FAQ answer. It’s an article, and it should be structured as one.

Add proper FAQPage schema. This is non-negotiable. An FAQ section without schema markup is just text. An FAQ section with schema markup is a structured data source that AI systems can parse programmatically.

Place FAQ sections on relevant pages, not just a standalone FAQ page. Your roofing cost page should have FAQ schema with roofing cost questions. Your personal injury page should have FAQ schema with personal injury questions. Topical relevance matters.

Update quarterly. New questions come up as the world changes. If your FAQ hasn’t been updated in a year, it’s probably missing the questions people are asking right now.

Step 6: Monitor Your AI Visibility

You can’t improve what you don’t measure. And right now, most businesses have zero visibility into whether AI systems are citing them.

The monitoring landscape for AEO is still early. There isn’t a single tool that gives you a complete picture (and anyone selling you one is exaggerating). But here’s what you can do right now:

Track referral traffic from AI sources. In Google Analytics 4, check your referral traffic for sources like chat.openai.com, perplexity.ai, and claude.ai. I’ve put together a step-by-step guide for setting up a custom AI Tools channel group in GA4 so you can track this automatically. You can grab it on this site.

One more thing about monitoring, and this is something that drives me crazy. You can set up perfect AI referral tracking in GA4 and still see almost nothing because your cookie consent banner is blocking the analytics script before it loads.

I’ve seen this on dozens of client sites. The compliance tools are often configured to their strictest settings by default, blocking virtually everything, even though most analytics can be configured to collect the data you need without exposing any personally identifiable information.

Before you spend time building custom channel groups and dashboards, check your consent settings first. If your compliance tool is set too strictly, you’re not measuring anything useful, AI traffic or otherwise. It’s like installing a security camera and leaving the lens cap on.

Here’s a real example. Many automobile dealerships in the country have added a compliance banner to their website. ComplyAuto is one of the most common providers. The problem is that these banners are often configured too strictly, blocking analytics tracking that doesn’t expose any meaningful information.

The dealership thinks they’re being responsible. What they’re actually doing is flying blind. They can’t see where their traffic is coming from, they can’t measure what’s working, and they definitely can’t track whether AI systems are sending them customers. And these are businesses spending tens of thousands a month on digital marketing with no way to measure half of it.

Monitor Google Search Console for AI Overview appearances. Google now includes AI Overview impressions in the Search Results performance report. It’s still not a dedicated report, but there’s more data there than most people realize.

Track your schema markup health. Use Google’s Rich Results Test and Search Console’s structured data reports to make sure your schema markup is valid and error-free. Broken schema is worse than no schema because it can actively confuse AI systems.

Use the new AI visibility tools. Six months ago, I would have told you the measurement tools for AEO were basically nonexistent. That’s changed faster than I expected. Here’s what’s now available:

Semrush AI Visibility Toolkit (launched October 2025) is the most comprehensive option if you’re already in the Semrush ecosystem. It tracks brand mentions and visibility across ChatGPT, Google AI Overviews, AI Mode, Perplexity, and Gemini. The Prompt Research feature shows you the exact prompts users are typing that trigger citations. The Competitor Research report shows where competitors get mentioned and you don’t. If you’re going to invest in one paid AEO monitoring tool, this is the one I’d start with.

Ahrefs Site Explorer now includes an AI Overview filter that shows which of your keywords are triggering Google’s AI summaries and whether your site is being cited in them. Again, if you’re already in Ahrefs, you have this.

Ahrefs Brand Radar (this was previously called Brand Explorer, so update your bookmarks) monitors AI mentions across six AI platforms with over 190 million prompts, tracking your brand’s visibility, sentiment, and citations within AI-generated responses.

Swiftly is specifically designed to monitor AI mentions, tracking your brand’s presence, share of voice, and competitor performance across major generative AI platforms like ChatGPT, Gemini, Claude, and Google AI Overviews, helping brands optimize for these new search environments.

LLM Pulse is specifically designed to monitor AI mentions, tracking how brands appear in responses from large language models like ChatGPT, Gemini, and Perplexity, analyzing sentiment, citations, and visibility to help businesses understand their AI search presence and competitive standing.

LLMRefs is a newer, specialized tool where you enter your domain and see if it’s being referenced across multiple AI platforms. I’m still evaluating this one, but the concept is exactly what the industry needs.

Google Search Console still doesn’t have a dedicated AI citation report, but you can spot AI-driven behavior by looking for queries with high impressions but unusually low click-through rates. That pattern often means an AI summary answered the question using your page as a source, but the user never needed to click through. The Search Results performance report now includes AI Overview impressions, so there’s more signal here than there used to be.

OLMoTrace is worth knowing about if you want to go deep. It uses string matching to trace whether your content was used in an AI model’s training data, not just cited in real-time answers. That’s a different question, but an important one.

I’ll be testing as many of these across real client campaigns as I can. I plan to publish detailed reviews with real data somewhere this site as I learn what’s actually worth paying for and what’s just a shiny dashboard. Check back or join the email list if you want those updates as they come.

One More Thing: llms.txt

There’s a newer standard worth knowing about called llms.txt. Proposed by Jeremy Howard of Answer.AI in 2024, it’s a markdown-formatted file you place at your domain root (yoursite.com/llms.txt) that tells AI systems where your highest-value content lives. Think of it as a curated sitemap specifically for language models.

If robots.txt is the bouncer that tells bots where they can’t go, llms.txt is the concierge that points them toward the good stuff.

Here’s my take: the concept is smart and the implementation is dead simple. But as of now, no major LLM lab has officially committed to reading it. An August 2025 audit of CDN logs across 1,000 domains found zero requests from GPTBot, ClaudeBot, or PerplexityBot for the file. That said, Yoast has already built automated llms.txt generation into their WordPress plugin, and the trajectory feels similar to where robots.txt was in its early days.

If you want to be ahead of the curve, add one. It takes 15 minutes. If you want to wait for proof that it matters, that’s reasonable too. I’ve added one to this site because the cost is basically nothing and the potential upside is worth the bet.

The Mistakes I See Over and Over

I’ve seen AEO strategies implemented across competitive industries for a while now, and I keep seeing the same mistakes. Some of them are technical. Most of them are strategic. All of them are fixable.

Mistake #1: Treating AEO as a separate project from SEO. AEO isn’t a new discipline that replaces your existing work. It’s an extension of good SEO practices. If someone tells you to “pivot to AEO” and abandon your current SEO strategy, they’re wrong. The best AEO work builds on a strong SEO foundation.

Mistake #2: Creating content specifically “for AI” that nobody would actually want to read. I’ve seen sites create robotic, keyword-stuffed FAQ pages that technically have the right schema but read like they were written by a committee of machines. AI systems are trained on human content. Write for humans first. Structure it for AI second.

Mistake #3: Ignoring E-E-A-T because you can’t put it in a spreadsheet. I get it. Schema markup is a checklist item. You either added it or you didn’t. Building author authority feels like something you can’t measure, so it gets pushed to the bottom of the to-do list. There are sites with perfect technical implementation getting passed over by AI systems in favor of sites with strong, verifiable expertise. The AI is looking for the real expert. Make sure it can tell that’s you.

Mistake #4: Setting it and forgetting it. AEO is not a one-time project. AI models update. Search behavior changes. Your competitors catch up. The businesses that treat AEO as an ongoing practice (updating content, adding new schema, refreshing FAQ sections) are the ones that maintain their citation rates. The ones who “did AEO” six months ago and moved on are already losing ground.

Mistake #5: Not monitoring results. If you’re not checking whether AI systems are actually citing you, you have no idea if your AEO work is paying off. Tools like Semrush’s AI Visibility Toolkit, Ahrefs Brand Radar and others are making it easy to track this automatically. If monitoring your AI presence still sounds like too much effort for something that could determine whether your business shows up in the future of search, recalibrate your priorities.

Mistake #6: Blocking the AI Bots You’re Trying to Impress. Many businesses are running Cloudflare or similar CDN services that block AI crawlers by default. GPTBot, ClaudeBot, PerplexityBot, they’re all getting turned away at the door. And it’s not just CDN settings. Overly aggressive compliance banners and bot protection tools can block AI crawlers the same way they block analytics scripts. Your content could be perfectly structured, your schema could be flawless, and none of it matters if the AI can’t read your site in the first place. Check your CDN settings, your robots.txt file, and your compliance tools. If you’re blocking AI crawlers, you’re doing AEO with the front door locked.

Mistake #7: Only optimizing your website and ignoring the rest of the internet. This is the new one, and it might be the most important. AI systems don’t just evaluate your site. They evaluate what the entire web says about you. If your website is perfectly optimized but nobody mentions your brand on Reddit, LinkedIn, YouTube, or in industry publications, you’re missing the signals that ChatGPT and Perplexity use to determine who’s worth citing. Build your authority everywhere, not just on your own domain.

Frequently Asked Questions About AEO

What is Answer Engine Optimization (AEO)?

AEO is the practice of structuring your website’s content so that AI systems like ChatGPT, Google AI Overviews, and Perplexity can find your information, understand it, and cite it when generating answers to user queries. It’s an evolution of traditional SEO that focuses on being the source AI recommends, not just ranking in search results.

Is AEO replacing SEO?

No. AEO builds on top of a strong SEO foundation. You still need good content, technical SEO, and genuine authority. AEO adds a layer of optimization specifically for how AI systems discover, process, and cite your content. Abandoning SEO for AEO is a mistake. Ignoring AEO while only doing traditional SEO is also a mistake. Google’s John Mueller put it clearly in December 2025: there is no AEO or GEO without SEO fundamentals.

How long does AEO take to show results?

It depends on your starting point. Sites with strong existing SEO (good content, solid authority, clean technical setup) can see AI citations within weeks of implementing schema markup and content restructuring. Sites starting from scratch need 3-6 months to build enough authority. There’s no shortcut for genuine expertise.

What’s the difference between AEO and GEO?

The industry hasn’t fully sorted this one out yet, and anyone who tells you there’s a clean, settled definition is getting ahead of the consensus.

AEO (Answer Engine Optimization) generally refers to optimizing for any system that delivers direct answers, including Google’s featured snippets, voice assistants, and AI platforms like ChatGPT and Perplexity.

GEO (Generative Engine Optimization) is a newer term that specifically focuses on AI systems that generate synthesized answers from multiple sources, like Google AI Overviews, ChatGPT, and Perplexity. Andreessen Horowitz published a major thesis using the GEO term in May 2025, and it’s gotten traction. But there’s a camp (including Profound, a well-known AI search platform) that argues AEO is the better term because GEO conflicts with geography, geology, and geo-targeting. Wikipedia has a full article on GEO as the primary term. Meanwhile, Google’s John Mueller lumped them all together and basically said the name doesn’t matter as long as you’re doing the fundamentals.

My take: the strategies overlap almost entirely. The label matters less than the work. I use AEO because it’s clearer and more intuitive to business owners who already understand SEO. If I use GEO, some of them may think I’m talking about geo-targeting, geo-fencing, or a Chevy from the ’90s. See the GEO Tactics guide on this site for the full breakdown.

Do I need technical skills to implement AEO?

The content restructuring (putting answers first, building FAQ sections) doesn’t require technical skills. Schema markup implementation is slightly technical but manageable with plugins like Rank Math if you’re on WordPress. For custom schema or complex implementations, you’ll likely want a developer or an experienced SEO professional.

Which AI systems should I optimize for?

Focus on the big three: Google AI Overviews (because Google still dominates search), ChatGPT (800+ million weekly active users and growing), and Perplexity (the fastest-growing AI search platform, now handling hundreds of millions of queries per month). The good news is that the core AEO practices (structured content, schema markup, strong E-E-A-T, multi-platform authority) work across all of them. But be aware they don’t all work the same way behind the scenes, which is why building authority across the web matters, not just on your own site.

How much does AEO cost to implement?

If you’re doing it yourself with this guide, the cost is your time. If you’re using an agency or consultant, expect to invest similar amounts to what you’d pay for a comprehensive SEO campaign. The specific number depends on the size of your site, your industry, and how much existing content needs restructuring.

Key Research Sources Referenced in This Guide

Every claim in this guide is backed by a specific source. If you want to verify anything, here’s where to look. If a link is dead, let me know and I’ll fix it. Broken citations on a site about credibility would be embarrassing, and I’d deserve to hear about it.

AI Search Impact and Click-Through Rate Research

Seer Interactive: AIO Impact on Google CTR (September 2025 Update)
Analysis of 3,119 informational queries across 42 organizations tracking 25.1 million organic impressions. Found 61% organic CTR drop and 35% more clicks for cited brands. The primary data source for the click-through rate impact claims throughout this guide.

Pew Research Center: Google AI Summaries and User Click Behavior
Behavioral tracking study of 900 U.S. adults analyzing 68,879 Google searches in March 2025. Found only 8% of users clicked a result when AI summaries were present (vs. 15% without), and just 1% clicked citations within summaries.

Similarweb: Zero-Click Search and AI Overview Impact Report (July 2025)
Tracking data showing zero-click searches grew from 56% to 69% between May 2024 and May 2025. Organic traffic to news publishers dropped from over 2.3 billion visits to under 1.7 billion in the same period.

seoClarity: The Overlap Between AI Overviews and Organic Rankings
Analysis of 432,000 desktop keywords showing 97% of AI Overviews cited at least one source from the top 20 organic results. Position 1 pages appeared in AI Overviews more than half the time.

AI Citation Patterns and LLM Visibility Research

Search Atlas: The Limits of Schema Markup for AI Search
December 2025 study analyzing schema markup coverage against LLM citation rates across OpenAI, Gemini, and Perplexity. Found that domains with extensive schema were cited no more frequently than domains with little or none, indicating schema alone does not influence LLM source selection.

The Digital Bloom: 2025 AI Citation and LLM Visibility Report
Comprehensive analysis synthesizing academic research and proprietary industry data analyzing 680 million+ citations. Found that brand search volume (not backlinks) is the strongest predictor of AI citations, with a 0.334 correlation. The source for claims about brand authority being 3x more important than backlinks for AI visibility.

Search Atlas: Comparative Analysis of LLM Citation Behavior
Study evaluating citation behavior across 5.5 million responses from 748,425 queries collected between August and September 2025, covering Gemini, OpenAI, and Perplexity. Reveals structural differences in how each platform retrieves and attributes sources.

RAG and Content Retrieval Research

Ahrefs: The AI Bots That ~140 Million Websites Block the Most
Analysis of approximately 140 million websites showing GPTBot is the most blocked AI crawler at 5.89% of all websites. ClaudeBot saw the highest growth in block rates, increasing 32.67% over the prior year.

Jeremy Howard / Answer.AI: The /llms.txt Proposal
The original September 2024 proposal for the llms.txt standard, a markdown file placed at the domain root to tell AI systems where a site’s highest-value content lives. Referenced in the guide’s discussion of llms.txt implementation.

Google Statements and Industry Analysis

Google Search Live (Zurich, December 2025): John Mueller on SEO Fundamentals. An industry post by Ben Kazinik details the Zurich event where Mueller stated: ‘AI systems rely on search, and there is no such thing as GEO or AEO without doing SEO fundamentals.’ Referenced as confirmation that AEO builds on traditional SEO, not replaces it.

What to Do Next

If you’ve read this far, you know more about AEO than 95% of your competitors. That’s not an exaggeration. Most businesses haven’t even started thinking about this.

Use the Keep Learning: Related Guides stack below to go much deeper into GEO, E-E-A-T, YMYL, and Schema Markup.

If you want a monthly update on what’s working: Join The Punch List monthly email newsletter. One email a month, no spam, genuinely useful.

And if you’ve got a question this guide didn’t answer, reach out. I read everything.

Keep Learning: Related Guides

→

The Complete GEO Guide

Practical strategies for getting cited by AI systems

→ →

E-E-A-T for AI Search

Building authority that AI systems actually recognize

→ →

YMYL Guide

Why AI holds your industry to a higher standard

→ →

Schema Markup for AI Search

What actually works for structured data (and what’s just noise)

→

The Punch List

If you want to stay connected, The Punch List lands once a month. What I learned, what I tested, what surprised me, and what I think it means for your business. No emoji. No “growth hacks.” Just the useful stuff, from someone who’s actually doing the work.

One email a month. Real observations, not recycled advice. Unsubscribe anytime.