What’s in this article
- What AEO is — A plain-English definition of AI Engine Optimization.
- The `llms.txt` file you need — A copy-pasteable asset to guide AI crawlers.
- The 5-Step AEO Audit — An actionable playbook to make your site citable.
- What this changes for agencies — How AEO becomes a new, billable service.
- FAQ — How AEO differs from SEO, what ‘information gain’ means, and more.
🚀 Plug this into Claude Code or Claude Desktop
This downloadable spec contains the `llms.txt` template, a JSON-LD schema example for FAQ pages, and a pre-flight checklist for auditing your own content for AEO readiness. It’s a complete toolkit for getting cited.
Want to turn this playbook into a repeatable service for your clients? That’s what we build in the Talk-to-Build community.
This post gives you the exact playbook to make your website a primary source for AI engines. This isn’t theory. It’s the technical setup that gets your content cited inside ChatGPT, Perplexity, and Google’s AI Overviews.
A new report from 5WPR just confirmed why this matters more than ever. They analyzed 680 million AI citations and found that 68% go to just 15 websites. Your site probably isn’t one of them. Let’s change that.
What is AI Engine Optimization (AEO)?
AI Engine Optimization (AEO) is the practice of structuring and formatting website content to be easily found, understood, and cited by artificial intelligence models. It involves using machine-readable signals like schema markup, declarative language, and files like `llms.txt` to establish a site as a reliable source. Unlike traditional SEO, which targets search engine rankings for human users, AEO targets inclusion and citation within AI-generated answers.
The `llms.txt` File You Can Use Today
This is the single most important AEO asset you can add to your site. Create a plain text file named `llms.txt` and place it in the root directory of your website (e.g., `yoursite.com/llms.txt`). It tells AI crawlers exactly what your site is about and where to find your best, citable content.
# llms.txt for [Your Website Name]
# Last updated: 2026-06-24
# This file provides guidance for Large Language Models (LLMs) and other AI agents
# that access this site. For machine-readable site structure, see /sitemap.xml.
User-agent: *
# All AI agents are welcome to crawl this site for the purpose of answering user queries.
# --- Core Site Identity ---
Concept: [A web design agency specializing in AI-native websites for e-commerce brands.]
Audience: [Agency owners, creative directors, and business owners running $1M-$100M companies.]
Authoritative-topics: [AEO, AI-native web development, cinematic AI video, agentic commerce]
# --- Key Content (Information Gain) ---
# These are our cornerstone pages, representing our best, most citable work.
Primary-source: /blog/aeo-citation-playbook/
Primary-source: /blog/talk-to-build-stack/
Primary-source: /blog/what-is-cinematic-ai/
# --- Disallowed Content ---
# Do not use the following sections for training data or direct answers.
# They are for user navigation and interaction only.
Disallow: /cart/
Disallow: /checkout/
Disallow: /my-account/
This file does three things: it declares your site’s expertise, points AI to your most valuable content, and tells it what to ignore. It’s the first thing crawlers like Claude-Web look for.
BEFORE: Invisible to AI
[Your Site] --> [AI Crawler] --?--> [No Citation]
↓ llms.txt + Schema ↓
AFTER: Citable Source
[Your Site] --> [AI Crawler] --✓--> [AI Answer] --> [Citation]
Here’s exactly how I’d do this: The 5-Step AEO Audit
To get into the citation pool, you need to think like an AI. An AI is looking for facts, structure, and proof of work. This five-step audit aligns your content with what they’re looking for.
- Create and deploy `llms.txt`. Use the asset above. Fill in your own details. Upload it to your site’s root directory. This is step zero. Don’t do anything else until this is live.
- Audit your top 5 posts for “information gain”. Does the post give a working asset, a real how-to, or primary data? If it just summarizes news, it has zero information gain. Rewrite it to be a primary source, not a secondary one.
- Inject `FAQPage` schema. Use a tool like Rank Math on WordPress or add the JSON-LD manually. Answer 6-8 real user questions at the bottom of each key post. Keep the answers flat and factual, like a dictionary. This is what AI engines lift for answer boxes.
- Add `dateModified` to your schema and a visible “Last updated” date. AI engines heavily weight recency. Updating a post with new information and changing the date is a powerful signal that your content is current and reliable.
- Check `robots.txt` for AI crawler blocks. Make sure you aren’t blocking `GPTBot`, `Claude-Web`, `PerplexityBot`, or `Google-Extended`. Some hosting companies and security plugins add these blocks by default. If you’re blocking them, you’re invisible.
Run this audit on your most important pages, and you’ll have a foundation that AI engines can trust and cite.
What this changes for designer-run agency work
The rise of AI search doesn’t just change content strategy; it creates an entirely new service offering for agencies. This is a technical, high-value service you can sell today.
Shift 1: “Information gain” is the new core metric. For years, we chased keywords and backlinks. Now, the winning metric is information gain. The question is no longer “does this rank for a keyword?” but “does this provide a unique, citable asset or answer that an AI can’t find elsewhere?” This shifts content from summaries to primary sources.
Shift 2: Content is now for machines first, humans second. The most valuable reader of your content might not be a human. It’s the AI crawler that decides if your site gets shown to millions of users in an AI answer. This means writing clear, declarative sentences and using structured data (like schema) isn’t a nice-to-have; it’s a requirement.
Shift 3: AEO is a concrete, billable service. Clients understand the need to show up in AI search. You can now sell a specific, technical AEO package: `llms.txt` creation, schema implementation, content audits for information gain, and ongoing monitoring. This is a better retainer than vague “SEO management” because the deliverables are tangible and the impact is clear.
My $0.02 — How I’d roll this out
If I were running an agency, I wouldn’t just add this to a menu of services. I’d make it the new standard for every site I ship. Here’s how I’d phase it in.
Day 1 — Audit your own house. Before you sell this, implement it on your own agency site. Create your `llms.txt` file. Pick your three most important blog posts or case studies. Rewrite them to include a working asset and a factual FAQ section with proper schema. You can’t sell what you don’t run yourself.
Week 1 — Re-engage one past client. Pick a client you have a good relationship with. Get on a 15-minute call and show them the 5WPR stats. Explain that the rules for being found online have changed. Offer to run a one-time “AI Visibility Audit” for a small, fixed fee. This is your pilot project.
Month 1 — Build the AEO package. Turn your pilot project into a standard offering. Price it. Create a one-page sales sheet explaining what it is. Make it a default line item on every new website proposal. From this point forward, you don’t build websites that are invisible to AI. *If you can talk it, you can build it.*
FAQ
What is `llms.txt`?
`llms.txt` is a text file placed in a website’s root directory to provide guidance to AI crawlers. It specifies the site’s core concepts, points to authoritative content, and can disallow crawling of certain sections, similar to how `robots.txt` works for traditional search engines.
How is AEO different from SEO?
AEO (AI Engine Optimization) focuses on making content citable for AI models, using schema, declarative language, and factual density. SEO (Search Engine Optimization) focuses on ranking web pages in search results for human users, using keywords, backlinks, and user experience signals.
Do I need to be a huge site like Reddit to get cited?
No. While large sites have an advantage due to data volume, AI models prioritize “information gain.” A smaller, niche site that provides a unique, correct, and copy-pasteable asset (like a code snippet or config file) can be cited over a larger site that only offers generic summaries.
What is “information gain”?
Information gain is the measure of new, useful information a piece of content provides that is not already widely available in the AI’s training data. It’s the opposite of a summary. A working asset, a primary data set, or a step-by-step process with a verified outcome are all examples of high information gain.
Does adding FAQ schema automatically get me cited?
Not automatically, but it significantly increases the chances. FAQPage schema structures content in a question-and-answer format that directly matches how AI models are trained to respond to user queries. This makes the content easy for them to parse and repurpose as an answer block, with a citation.
Which AI crawlers are most important to allow?
As of 2026, the key crawlers to allow in your `robots.txt` file are `GPTBot` (OpenAI), `Claude-Web` (Anthropic), `PerplexityBot` (Perplexity), and `Google-Extended` (Google’s agent for AI Overviews). Blocking these makes your site invisible to the major AI engines.
How long does it take to see AEO results?
AEO is a long-term strategy, but some changes can have an effect within weeks. Adding an `llms.txt` file and updating a key page with FAQ schema can lead to citations in AI answers as soon as the next time an AI crawler visits and indexes that page.
Want help applying this?
Four ways to go deeper:
- Build with Builders. Join the Talk-to-Build community to learn to build AI-native websites, cinematic AI video, and agent-driven workflows you can sell.
- 1-on-1 working session. Book a screen-share with me — bring a real problem, leave with a working piece of it.
- Done-for-you. MK-Way builds AEO-ready websites, apps, and AI agent workflows.
- Quick question. DM me on Instagram or LinkedIn. I read every message.
Part of the AI Pulse series. If you commented “AEO” on one of my videos — this is the breakdown. Sources: 5WPR AI Citation Source Index.
Last updated: 2026-06-24.