AEO Best Practices That Don't Need a Big Budget

The best AEO best practices cost nothing except your time. Answer-first content, clean HTML, and a robots.txt policy that treats AI crawlers as three different problems, not one. That's the whole budget.

Search is splitting in two. Google still sends people to your site. But a growing number of visitors never touch a search engine. They ask ChatGPT, Claude, or Perplexity directly, and those assistants decide which sources to cite. If your site isn't structured for that, you're invisible in a channel that didn't exist five years ago.

The good news: you don't need enterprise software or an SEO agency to fix it. You need to understand how AI discovery works and make your pages easy for it. Here's the playbook.

Know which AI crawlers matter

AI discovery isn't one thing. It's three things, and each one behaves differently. Kymo's crawler registry classifies them into three categories, and it's worth learning the difference because your robots.txt strategy depends on it.

Training crawlers like GPTBot (OpenAI's training crawler) and ClaudeBot bulk-download pages to build foundation models. They send no traffic back to you. Blocking them is reasonable, and it won't remove your site from AI answers because the operators use separate agents for retrieval.

Indexing crawlers like OAI-SearchBot (OpenAI's search index crawler), PerplexityBot, and Claude-SearchBot work ahead of time, like Googlebot. They build a search index that later answers pull from. These can send referral traffic, and you generally want to let them in.

Live fetchers like ChatGPT-User (OpenAI's live fetcher), Claude-User, and Perplexity-User make a fetch when one person asks a question right now. These are your highest-intent visitors. They can send referral traffic, but they're also the hardest to optimize for because you can't predict exactly what triggered them.

Here's the table that matters:

One note: Perplexity has no training crawler. It builds no foundation models. Both of its agents handle retrieval, PerplexityBot indexes, Perplexity-User fetches live. There is no third agent. And Claude-SearchBot is the indexer, Claude-User is the live fetcher. Published directories often reverse these two. Don't.

For a fuller breakdown of which crawlers to allow and which to block, read Should You Block AI Crawlers? (It Depends Which Kind).

Answer the question directly, early, and in plain language

AI assistants prefer sources that are easy to quote. That means your content should answer a question in the first paragraph, not the fourth. The pattern is simple: state the answer, then explain it. This is opposite to how most marketing blogs are written, where the answer gets buried under context, statistics, and a story.

You want a clean sentence that works as a citation. "AEO best practices that don't need a big budget include writing concise answers, keeping pages crawlable, and tracking AI crawler behavior." That sentence alone is quotable.

The free Kymo's free AI visibility checker can help you see how your pages currently look to AI agents. It reads a URL from public signals only, so your site doesn't need Kymo installed, and the report arrives by email. It's a zero-cost way to find out if an assistant can extract a clean answer from your content.

Structure your pages for extraction

AI crawlers parse HTML the way search engines did in 2010. They like semantic structure. H1 for the page title, H2s for sections, short paragraphs, and list markup where it fits. This isn't glamorous, but it works.

Some specifics:

• One H1 per page. Make it match the question you're answering.
• Put the key answer in the first 100 words.
• Use tables for comparison content. Assistants can cite tabular data well.
• Keep paragraphs under four lines in most cases.
• Avoid paywalls and login walls on content you want AI agents to see.

If your site is a JavaScript single-page app, you have a problem. Many AI crawlers render JavaScript poorly or not at all. Serve static HTML or use server-side rendering if you want indexing crawlers to read you.

Optimize for the AEO flywheel

AEO has a positive feedback loop. An indexing crawler reads your page and stores it. A user asks a question. The assistant cites your page. The user clicks through. That click is a referral, which you can see in your analytics.

The thing is, you can't improve this loop if you can't see it. Google Analytics won't show you AI citations, because it tracks humans, not crawlers. Why GA4 Can't Show You AI Citations explains the gap in detail.

What you need is a dashboard that separates AI crawler activity from human traffic. The Live Kymo demo dashboard shows what that looks like. You can see which pages ChatGPT-User fetched, which ones PerplexityBot indexed, and which ones generated actual click-throughs. That data tells you what to write next, what to fix, and what to promote.

The The Small Business Guide to AEO walks through this process in order. Start there if you want the full playbook without the fluff.

Make your pages fast, clear, and quotable

Core web vitals matter for AI crawlers more than you'd think. Slow pages get crawled less. Pages with messy, duplicate content produce muddy answers. And pages that conflict with themselves, like having two different answers to the same question, get ignored or misquoted.

Practical, free steps:

1. Compress images. TinyPNG is free and gets most pages down to a reasonable size.
2. Remove render-blocking JavaScript from above the fold.
3. Use a caching plugin if you're on WordPress.
4. Make sure your meta descriptions actually describe the page. Assistants read them.
5. Check your robots.txt for accidental blocks.

For the last point, remember that control tokens like Google-Extended and Applebot-Extended are not crawlers. They're robots.txt tokens only, and they don't have user agents. Disallowing one controls how content an operator already crawled may be used. It never appears in a log and it never "visits" your site.

Track AI visibility before you optimize

You can't improve what you don't measure. That's the core argument, and it's especially true for AEO because the mechanics are still new and the crawler behavior changes frequently.

Kymo is a hosted product, not self-hosted, and it's simple enough to check. Both plans include AI visibility tracking. The Solo plan is $9 a month or $90 a year, up to 10,000 events per month. The Studio plan is $29 a month or $290 a year, up to 100,000 events per month. Every feature is included on both plans, they differ only by event volume. Both start with 14 days free, no card required.

AI crawler tracking does not count against your event limit. So a site with lots of bot traffic won't burn through its quota.

The The Best AI Visibility Tools, and What Each One Can't Tell You post compares Kymo against other options. The short version: single-purpose AI trackers give you crawler data but no human analytics, and broad SEO suites like Semrush and Ahrefs are priced for agencies. Kymo sits in the middle, one dashboard for humans and AI agents, built for solo builders.

Stop worrying about the wrong metrics

Traffic from human visitors is still your primary goal. AI referrals are a bonus channel, and for some sites they're becoming a major one. But obsessing over citation counts without knowing whether citations turn into visits is a mistake.

Here's what to track:

• Which pages AI crawlers fetch most.
• Which prompts likely triggered those fetches.
• Which AI referrals actually clicked through.
• Your overall human visitor trend over time.

When you exceed your event limit on Kymo, the dashboard pauses but your data is not deleted. That's a nice safety net, but the free 10,000 events a month is enough for most solo builders.

Write for humans first, AI second

The overlap between good content for humans and good content for AI assistants is enormous. Clear structure, direct answers, and useful information. If you write something genuinely helpful, you're 80% of the way there.

The remaining 20% is technical. Formatting your answers so a crawler can extract them. Keeping your pages fast. Making sure your robots.txt doesn't accidentally block the indexing crawlers you want. Checking your analytics to see what's working.

Both Solo and Studio start with a 14-day free trial, no card required, and AI crawler tracking doesn't count against your event limit. That's enough time to see which AI agents are already visiting your site, what they're fetching, and whether any of it sends people back.

Measure what you're already getting for free

The cheapest AEO practice is to look at your current logs. Most hosting providers give you raw access logs for free. Search for "GPTBot", "PerplexityBot", "ChatGPT-User", and "ClaudeBot" and you'll see what's been crawling. That's a manual version of what Kymo automates, and it costs nothing.

But raw logs are hard to read, and they don't tell you which pages matter or whether crawls led to clicks. A tool that classifies and visualizes this data saves you hours a week. That's where Kymo comes in.

The one practice that matters most

Every article about AEO lists tactics. The realistic advice is simpler: pick a few pages that answer real questions, make those pages impeccably structured, allow indexing crawlers, and measure the result.

That's it.

You don't need a content team. You don't need a $200-a-month SEO suite. You need discipline and you need visibility.

Start with the guide, then start tracking

Work through the The Small Business Guide to AEO to audit your content structure, set up your robots.txt policy, and write answers that AI assistants can cite. Then use Kymo's free AI visibility check to see how a URL looks before you install anything. It reads the page from public signals only, so the site does not need Kymo installed, and the report arrives by email. After that, start the trial, track what the crawlers actually do, and iterate from data instead of guesses.

Start free → 14-day free trial. No card required.