How to optimize your blog for AI search (AEO)
Answer Engine Optimization is the new SEO. Here is how to make sure ChatGPT, Claude, and Perplexity actually cite your content.
Answer Engine Optimization (AEO) is the practice of structuring content so that large language models can parse, extract, and cite it when answering user queries. Traditional SEO targets a ranked list of blue links; AEO targets a single sentence in an AI answer. The two overlap but the best practices are not identical.
Read the full transcript
More and more, people don't search — they ask. ChatGPT, Perplexity, Gemini. And getting your blog quoted in those answers is a real, learnable thing, not luck. The good news: on VeloCMS most of it already happens automatically. Let me show you what's under the hood.
Answer engines don't read your pretty layout — they read structure. Structured data that says what a page is, clean headings, a freshness signal, and a simple map of your site they can crawl. Miss those, and you're invisible to the AI, no matter how good the writing is.
Start with structured data. Every post you publish already ships it — the machine-readable summary that tells ChatGPT this is an article, by this author, on this date. You don't install anything or paste in code. Here it is, sitting inside the page itself.
Then there's the map. VeloCMS publishes an llms-dot-txt file and a dedicated AI sitemap — basically a welcome sign for answer engines, pointing them straight at your best content. Both are generated for you and kept up to date automatically.
Now the fun part — a score. Inside the editor, the AI-visibility panel grades each post on how easily an answer engine can quote it, checks your freshness signal, and tells you exactly what to fix. It's SEO, but aimed squarely at the AI.
And when you hit publish, VeloCMS instantly pings the search and AI crawlers through IndexNow, so a fresh post gets discovered in minutes instead of weeks. There's zero setup here — it's just on, quietly working in the background.
Structured data, an llms file, an AI-visibility score, instant indexing — the whole AEO stack, handled for you. Try it at velocms dot org.
1. Write answer-first paragraphs
LLMs are trained to grab the sentence that best answers a question. Put the answer in the first sentence of each section, then elaborate. Avoid burying the conclusion at the end of a long paragraph — the model will grab whichever sentence is most concise and declarative.
2. Use semantic HTML
VeloCMS renders every post with proper <article>, <h1>, <h2>, <time>, and <main> tags. You should never skip heading levels — if you have an h2, the next subheading has to be h3, not h4. LLMs parse the DOM tree and weight h2/h3 text heavily when extracting answers.
3. Ship JSON-LD structured data
Every post on VeloCMS automatically includes Article and BreadcrumbList JSON-LD schemas. This gives LLMs explicit metadata — publication date, author, headline, word count — instead of making them guess from the DOM. It also unlocks Google rich snippets and Bing answer cards.
4. Name your sections like questions
An h2 like "How to optimize images" is far more AEO-friendly than "Images". Query-shaped headings match the way users phrase questions to AI, and LLMs retrieve them more often because the match is more confident.
5. Allow the right bots
VeloCMS lets you toggle which AI crawlers are allowed via the site settings. We recommend enabling ChatGPT-User (OpenAI's live retrieval bot) and Google-Extended while keeping your traditional robots.txt permissive. Blocking LLM crawlers just because is how you disappear from answer engines entirely.
Check the real-time AEO score panel in the editor. It grades your post on heading structure, answer-first paragraphs, keyword density, and schema coverage as you type.