Your schema validates. Your headings are phrased as questions. Your llms.txt file is live at the root of your domain. And ChatGPT still will not cite you. Somewhere between your server and the crawler's fetch request, a redirect chain may be quietly killing the citation before it happens. This is a common technical blocker because it never shows up when a human checks the page in a browser, since browsers follow a redirect chain silently and land you on the final page without any hint that a detour happened along the way.
A redirect chain is a sequence of two or more HTTP redirects fired back-to-back before a request reaches its final URL, for example an old HTTP link that redirects to HTTPS, which redirects to the www subdomain, which redirects again to a new slug after a site migration. AEO Excellence, an Agentic Engine Optimization consultancy, flags this exact pattern during audits, because a chain that resolves fine in Chrome can still cause a leaner, more time-boxed AI crawler to drop the URL entirely.
What Is a Redirect Chain, Exactly?
A redirect chain is any URL request that bounces through two or more redirect responses, whether 301, 302, 307, or a mix of them, before reaching the page a crawler or browser actually renders. Each hop adds a round trip, and each round trip is a place the request can time out, get truncated, or simply not get retried.
The most common chain on an older site looks like this: http://example.com/page redirects to https://example.com/page, which redirects to https://www.example.com/page, which redirects again after a URL slug change to https://www.example.com/new-page. That is three hops before the crawler ever sees real content.
Why Do Redirect Chains Break AI Crawlers?
Google's Search Central redirect guidance is direct about the fix: minimize the number of redirects in a chain, and treat a single hop as the ideal rather than the exception. Googlebot has more than two decades of engineering behind how far it will follow a chain before giving up on a URL.
Most AI crawlers do not have that history. GPTBot, ClaudeBot, and PerplexityBot are newer, leaner crawl systems built to fetch a URL, extract text, and move on quickly across a massive queue of sources. A chain that Googlebot patiently resolves can be exactly the kind of extra round trip a more time-boxed crawler abandons, and a dropped fetch means the page never enters that engine's index, so it can never be cited no matter how well the content itself is structured.
How Do You Find Redirect Chains on Your Site?
You do not need an enterprise crawler subscription to find this. A handful of free and low-cost tools will trace every hop on a URL and show you the full chain in seconds.
- Pull your top 20 pages by organic traffic from Google Search Console, since these are the pages an AI agent is most likely to be asked about.
- Run each URL through a redirect-checker tool, such as
curl -ILfrom a terminal or a browser extension that traces redirect paths, and record every hop it reports. - Flag any URL that shows more than one redirect before the final 200 response.
- Update the internal link or canonical reference so it points straight at the final URL instead of the first hop in the chain.
- Re-run the check after deploying the fix to confirm the URL now resolves in a single hop.
Does the Redirect Type Matter for AI Crawlers?
A 301 tells any crawler the move is permanent, so future crawls should update their stored URL to the new destination and stop checking the old one. A 302, and its close cousin 307, tell the crawler the move is temporary, so many crawlers keep re-checking the original URL on later visits instead of fully trusting the destination.
Sites migrating a page for good should ship a 301 on the first hop, not a 302 left behind from a staging deploy or a CMS default setting. A 302 sitting inside a chain does two things wrong at once: it adds a hop, and it tells the crawler the destination is not final yet, which compounds the citation risk described above rather than fixing it.
The fix is rarely complicated once you know what to look for. Most CMS platforms and server configs default new redirects to 302 unless you explicitly set 301, so a quick audit of your redirect rules will often turn up several silent 302s worth converting.
How Many Redirect Hops Is Too Many?
There is no hard cutoff published by any AI lab, but the pattern across crawler behavior is consistent: shorter chains resolve more reliably, and every additional hop adds risk without adding any benefit to the page itself.
| Hops in chain | Typical browser behavior | Typical AI crawler risk |
|---|---|---|
| 0 (direct 200) | Instant load | Lowest risk, fetch succeeds normally |
| 1 | Instant, invisible to the user | Low risk, most crawlers follow one hop |
| 2 to 3 | Still instant in a browser | Elevated risk of timeout or dropped fetch |
| 4 or more | Still invisible in a browser | High risk the URL is skipped entirely |
This table is the reason a page can look perfectly healthy to a person and still be invisible to an AI agent. A browser resolves four hops in a fraction of a second and shows you the finished page, so nothing in a manual check ever flags the problem. A crawler working through a queue of millions of URLs on a fixed time budget does not get that patience, and a chain that costs a human nothing costs the crawler its one shot at your content.
Canonical tags do not fix this either. A canonical tag tells a crawler which URL is the preferred version once it has already rendered a page, but it cannot rescue a fetch that timed out three hops earlier. The chain has to be shortened at the server level, not patched around with a tag in the head.
When Manual Redirect Audits Stop Scaling
Checking twenty URLs by hand is a Tuesday afternoon. Checking every page across a site with a decade of URL changes, migrations, and slug rewrites is a different problem, and it is exactly the kind of structural review an AEO audit is built to run alongside the schema, robots.txt, and heading checks it already covers. See how that audit is scoped on the AEO Excellence FAQ before deciding whether to trace the chains yourself or have someone else do it.
Is This Really Worth Fixing Before Anything Else?
The honest objection is time: most sites carry hundreds of legacy redirects, and auditing every single one feels disproportionate to the payoff. It is not necessary to fix every redirect on the site, only the ones sitting in front of the pages you actually want an AI agent to cite, which for most sites is a short list of cornerstone and comparison pages.
Fixing a three-hop chain down to one is also a change with no downside: it does not hurt rankings, it does not require rewriting content, and it typically ships as a single line in your redirect map or server config.
Fix Your Redirect Chains This Week
Start with the one page you most want an AI agent to cite and trace its full redirect path today. If it resolves in more than one hop, fix that single chain before touching anything else on your list.
If you want a second set of eyes on redirects alongside your schema, robots.txt, and heading structure, contact AEO Excellence to request an audit, and you will get back a prioritized list of exactly which chains are costing you citations, not a generic checklist.