Executive Summary
Technical SEO is not really the foundation of AEO. Rather, it is a cleanup layer, although necessary. The strategy is only an hour of work with no ongoing cost: fix your robots.txt allow list, ship static HTML, and lock down URLs once. The citations that actually move the needle (such as earned media, third-party mentions, and entity consistency) happen off your domain, where technical SEO tools can’t see or measure.
Introduction
Most AEO content tells you technical SEO is the foundation of Answer Engine Optimization. This story keeps getting told because SEO tool companies sell what they already measure. However, the research tells a different story:
Entity authority correlates with AI citations 4.8 times more strongly than technical SEO factors.
Earned media gets cited about five times more often than brand-owned content.
One analysis put earned media at 85.5 percent of all AI citations.
Those numbers do not leave room for technical SEO as the main lever. Rather, SEO is a cleanup layer: useful, but it is not the work that wins.
If you are making a budget decision right now
First, fund the three cleanup fixes once: robots.txt allow list, static HTML accessibility, and URL stability. Treat that as a one-time line item.
Then, move the rest of your AEO budget off-site: earned media, third-party citation cleanup, entity consistency, and analyst and podcast coverage.
If your current AEO vendor is selling you ongoing schema work, monthly Core Web Vitals reports, or AI-citation dashboards, you are funding the wrong layer.
Here is what technical SEO actually does for AEO, and what it does not.
The Three Real Overlaps
1. Access
AI crawlers need to reach your site, to be able to discover, mention, and cite you.
OpenAI runs three separate crawlers: GPTBot scrapes content for training future models, OAI-SearchBot fetches pages for ChatGPT search and citations, and ChatGPT-User makes on-demand fetches when a user or a Custom GPT requests a page.
Anthropic runs four: ClaudeBot fetches pages for chat citations and Claude-Web does broader web crawling. Claude-SearchBot handles retrieval and can be controlled separately from ClaudeBot, so you can permit retrieval while denying training.
PerplexityBot is the main Perplexity crawler.
CCBot is the Common Crawl spider that feeds the broader training data substrate for almost every other model.
These specific bots may not be the ones your robots.txt was written for. This is because most legacy allow lists were written before any of these existed. So, you may be blocking the bots that decide whether you appear in AI answers at all.
The minimum robots.txt allow list for AEO
User-agent: GPTBot Allow: / User-agent: OAI-SearchBot Allow: / User-agent: ChatGPT-User Allow: / User-agent: ClaudeBot Allow: / User-agent: Claude-SearchBot Allow: / User-agent: PerplexityBot Allow: / User-agent: CCBot Allow: /
Adjust per your training data policy. If you want to keep your content out of training but stay visible in citations, allow OAI-SearchBot and Claude-SearchBot while disallowing GPTBot and ClaudeBot.
Fixing this takes less than an hour and costs nothing. This method is the single highest-leverage technical AEO fix in 2026.
2. Static accessibility
AI crawlers handle JavaScript differently than Google does.
GPTBot, PerplexityBot, and the OpenAI search bots download the initial HTML response and move on. They do not execute JavaScript nor wait for components to mount. Furthermore, they do not retry.
ClaudeBot may execute some JavaScript, but the safest planning assumption is that none of them do.
This has a concrete cost. If your product page loads pricing from an API after render, the AI does not see the price. If your comparison table is built by a React component, the AI does not see the comparison. If your FAQ answers appear only after a “load more” click, the AI does not see the answers.
Whatever ships in your initial HTML payload is what these crawlers see. Server-side rendering or static HTML closes the gap. Single-page apps without SSR lose this entire surface.
3. URL stability
When an AI cites your site, it cites a specific URL. If you change that URL, the citation breaks. The AI then moves on to whoever fills the gap.
Clean, simple, lasting URLs are the one of the only technical AEO moves that pays off over time without ongoing work.
Try this in five minutes
Open ChatGPT, Claude, Gemini, and Perplexity. Paste this prompt into each one, swapping in your category and your brand name.
I am evaluating vendors in [your category]. Who are the top three or four companies I should consider, and what is each one known for? Cite your sources.
Read the citations. You will find a mix of your own content, competitor pages, outdated press releases, directory listings, and reviews from a version of your service you do not run anymore. That is your real retrieval layer: something your technical SEO audit cannot see any of.
What Technical SEO Does Not Do
Schema markup is not a top driver of AI citations.
It helps Google make rich snippets and gives crawlers a cleaner read of your entities. However, the earned media that drives most AI citations is news articles, podcast transcripts, Wikipedia pages, and Reddit threads. These sources have almost no schema.
Schema is still worth doing because it is cheap, but it just does not move your citation rates.
In addition, Core Web Vitals and page speed do not change AI citation rates either. This is because AI crawlers are not users: they do not bounce. So, a page that loads in 2.3 seconds gets cited the same as one that loads in 1.1. This site-speed obsession in most AEO content is leftover thinking from SEO.
Mobile-friendliness does not matter to most AI crawlers. They fetch raw HTML, thus the mobile view doesn’t matter. Mobile-first indexing is a Google rule, not an AI rule.
Internal linking and content clusters do not build entity authority with AI the way they build topic authority with Google. AI grabs single pages based on outside signals, and doesn’t really rely on your internal navigation. Internal linking still helps human readers and traditional SEO, but it does not move AEO.
HTTPS is the floor. With every real site having it, it does not help you stand out.
The Honest Split
Technical SEO is the cleanup layer of AEO. It keeps your site from being invisible, however, it is not where the win comes from. The win comes from off your site, living in earned media, Wikipedia entries, Wikidata pages, analyst reports, podcast appearances, third-party reviews, and the consistency of how your brand gets named across the open web.
None of that work happens in your CMS, nor does it up in a technical SEO report. Most agencies oversell the technical layer because the technical layer is what they have tools for. These may be dashboards for crawl errors, schema checks, and page speed.
On the contrary, the work that actually moves AI citations is harder to measure, harder to bill, and outside the tools they already own.
Where to Put Your Weight
Spend a small slice of your AEO budget on the three cleanup fixes above.
A summary
- Fix your robots.txt allow list this week
- Confirm that your core content ships as static HTML
- Lock down your URLs.
Then, spend the rest off-site:
- Get cited
- Clean up the wrong and outdated mentions of your brand that AI is currently using to describe you
- Build a steady entity footprint across the places AI platforms actually trust.
That is where you stop paying the Hallucination Tax. That is where you close the Dark Pipeline.
Three questions for any AEO vendor before you sign
1. Where does your data come from?
A real answer acknowledges that nobody has access to real user query data from the AI platforms, and explains the methodology used to approximate it. A weak answer claims proprietary insight without explaining the panel, the sampling, or the validation.
2. What is your process for cleaning up outdated third-party content?
A real answer walks you through how they audit, update, and remove inaccurate brand mentions across the open web. A weak answer skips this and pitches new content production.
3. How much of the work happens off our domain?
A real answer says the majority of it. A weak answer focuses on schema, your site, and your dashboard.
Want to know whether an AEO vendor is selling you the right thing? Read more about how to find out here.
The AIBAI Diagnostic
The AIBAI Diagnostic shows you the full picture that your brand sits inside. Going beyond your domain, this diagnostic identifies where the real gaps live and the layer that technical SEO cannot reach.
This article covered the cleanup layer. The work that actually moves AI citation rates is off-site, ongoing, and structurally outside the tools you already own. That is what the AIBAI Diagnostic maps and what we build against.