After a redesign or domain change, which old links might AI search stop citing?
After a site redesign or a domain change, old links that AI search previously cited won't necessarily break immediately, but they usually won't switch over automatically either. What decides the outcome is whether 301 permanent redirects cover the whole site, whether old and new pages correspond one-to-one, and whether the sitemap and internal links are updated in sync. Based on typical 2026 project delivery experience, the experience range for old citations being gradually replaced by new links is 2–8 weeks; if old URLs return 404 or soft 404, old links lingering in AI answers may point to dead links for a long time.
After a domain change or redesign, how do old links cited by AI search change?
AI answer engines don't read your site in real time; they rely on search engine indexes, retrieval-augmented snippets and cache. The citation links they give are usually the URL they “saw at the time.” After a redesign or domain change, if the old URL can still 301 to the new URL, the index will gradually update; if it is simply deleted, the old citation becomes a dead link.
Some teams assume that once the new site launches, AI will just know. In custom website development projects, it's common that search index updates and AI source refreshes are not in sync, so old links may keep appearing for weeks. If the new page's content drifts away from the old page's topic, AI may also stop citing that page even if the URL hasn't changed.
- Site-wide 301 coverage: old section pages and detail pages all having clear redirects is the precondition for keeping citation signals.
- One-to-one content correspondence: if the old page was about A, the new page should also be about A; if it's merged into B, citations get re-evaluated.
- Sitemap and canonical: pointing to the new domain and new URLs reduces confusion in index signals.
- External and internal links: once external citations and on-site links are updated, replacement usually happens faster.
How can you tell whether old cited links have already failed or been replaced?
You shouldn't judge from one AI platform alone, and you shouldn't draw conclusions from a single question. Check in four layers — index, redirects, content, AI citations — looking at different signals at each layer so you don't mistake cache lag for a failed migration.
- Check the old URL's status code: see whether it's 301, 404, or 200 with changed content; the common passing line is that all old URLs 301 to new URLs with a clear correspondence.
- Check sitemap and canonical: confirm the sitemap submitted under the new domain contains no old-domain URLs, and canonical points to the new URL rather than the old domain.
- Spot-check internal links and navigation: look for any remaining internal links to the old domain or old paths, especially in the top nav and breadcrumbs.
- Record AI citations across multiple platforms: ask core questions in several AI search tools and log citation link changes weekly; don't conclude from a single result.
What counts as passing: old URLs return 301, new pages cover the same topic, sitemap and canonical are updated, and the share of old links in AI citations drops week by week. If old links keep returning 404, the migration is only half done.
An actionable migration check framework: before, during, after
Treat a migration as taking inventory of citation assets, not as a pure technical swap. Check across three stages — before, during, and after migration — each with different owners and acceptance criteria, otherwise you may only discover the citation gap after launch.
- Before migration: crawl the old site's URL list, count pages with search traffic and AI citation records, and flag pages that need one-to-one correspondence.
- During migration: configure 301s one by one, sync the sitemap, fix canonical and internal links, and keep the old site accessible for a period.
- After migration: monitor old URL status codes, AI answer citation links and search index coverage weekly, and add redirects promptly when 404s appear.
Note: 301s should point to new pages with a clear correspondence, not everything to the homepage; for old pages with no matching content, keep an explanation page or 301 to the parent section rather than 404ing directly. In terms of experience range, checking site-wide 301s often takes 1–3 business days — it isn't something you finish on launch day.
A situation that comes up often on delivery projects: the redesign schedule is tight, so only the old homepage gets a 301, and section and detail pages never get checked. After launch, old section URLs 404 one after another, citation links in AI answers become dead links, and some search entry points are lost too. Reworking the redirects from the old URL list took nearly a week; in projects with many pages, this kind of rework has a typical range of 3–7 business days, depending on the number of old URLs and how fast external links are updated. That's the cost of treating migration as a technical swap without taking inventory of citations.
Keeping the original URLs, 301 to a new domain, or a full rebuild: how do the three compare?
Different migration approaches affect AI citations differently; don't judge only by how good the new site looks. Below is an experience-based comparison across cost, timeline and citation risk; specific projects will vary with page count and how much external citation exists.
- Redesign with original URLs kept: cost is typically lower, timeline roughly a few days to two weeks; citation risk is relatively small, but the template structure should stay as stable as possible — avoid major changes to heading hierarchy and body content paths.
- Domain change with site-wide 301s: medium cost, timeline typically 2–6 weeks for a fairly stable index replacement; medium citation risk, depending on 301 coverage and content correspondence.
- Full rebuild with all URLs changed: higher cost, timeline typically 4–12 weeks; higher citation risk — if old links keep no redirect, citations in AI answers may point to 404s for a long time.
If your site has very few AI citations, spending a large budget on citation migration protection isn't worth it; just make sure the new site is crawlable and the content is understandable. If the site already has stable organic traffic and AI citations, then 301s and content correspondence are must-haves.
When this applies and where it doesn't
This migration protection fits sites that already have organic search traffic, AI search citation records, or a stable source of inquiries. If a site just launched and has almost no external links or AI citations, the main task of a redesign or domain change is to get the new site crawled normally, not to rescue old citations.
- Fits: sites running more than a year, with section pages cited by AI, where a gap in inquiries after migration isn't acceptable.
- Fits: multilingual sites or sites with many product lines, deep URL hierarchies, and a large number of old links.
- Doesn't fit: small showcase sites with no search traffic and no AI citations — just launch the new site and submit the sitemap.
- Not necessary: keeping every old page alive indefinitely just to preserve citations, which can cause duplicate content and rising maintenance costs.
Boundary statement: AI search citations are not a stable contract; any migration can only lower the probability that old links fail, and cannot promise that old citations will be kept or will appear in AI answers immediately after migration.
Common pitfalls
- Only the homepage gets a 301, while section and detail pages still 404.
- The old site is shut down entirely, with no old domain resolution or redirect kept.
- Old and new pages cover different topics, so after the 301 AI re-evaluates and citations may not follow.
- The sitemap still submits the old domain and canonical points to the old URL, so index signals conflict with each other.
- After the redesign, large amounts of body copy are deleted, leaving only images or video, so AI has no citable paragraphs.
- Treating AI citations as rankings, and judging migration success by the result of a single prompt.
FAQ
Is it normal for old links to still appear in AI answers after a domain change?
Yes. Search index and AI source refreshes aren't in sync; the common 2026 experience range is gradual replacement over 2–8 weeks, during which old links and new citations may coexist.
If an old page has no new counterpart, should it just be deleted?
Direct 404s aren't recommended. You can 301 to a parent section or an explanation page with a clear correspondence, keep the access path, and then watch how citations change.
If a redesign keeps the original URLs, will AI citations update immediately?
Usually not immediately. When the URL stays the same, AI may still cite the old cache; if the main content and heading hierarchy stay stable, updates tend to be more predictable.
Should we do a separate round of GEO content for the migration?
You don't need to create separate content. Fix 301s, sitemap, canonical and internal links first, then add paragraphs to existing sections that can independently answer questions.
How long after migration can you judge whether it succeeded?
The typical observation window is 4–8 weeks. Old URL status codes, changes in AI citation links, and whether inquiry sources stay stable are more reliable than a single AI answer.
If you're about to redesign or change domains, first list the old URLs that have traffic and AI citation records, confirm the 301 target page and content correspondence one by one, and then launch. This applies to sites with accumulated citations; for brand-new small sites with no search traffic, just make sure the new site is crawlable and the content is readable — there's no need to overinvest in old citations.
-
Drone Accessories Company Website DevelopmentIncorporating gray as an accent with the pr ...
-
Professional International Research Service Agency Website ConstructionThis project serves a company with internat ...
-
The Construction of Group Websites for Asset Operation and Digital ServicesThis project is to create a website for a c ...
-
Thermal Test Equipment Website ConstructionFounded in 2022, this tech company focuses ...
-
At website delivery acceptance, which content might not get picked up by AI search later?
Date: Sep 11, 2026 Read: 6
-
A new service page has been live for a month, but AI search still hasn't cited it. What's the problem?
Date: Sep 10, 2026 Read: 11
-
Are custom websites more likely to be cited by AI search than template websites?
Date: Sep 9, 2026 Read: 15
-
In the Era of AI Search, How Can Custom Website Development Seize Traffic Opportunities with GEO?
Date: Jul 13, 2026 Read: 142
-
AI Search is Reshaping Website Traffic Patterns: Customized Content Becomes Key
Date: Jul 13, 2026 Read: 144




