Site migrations break things. You know this already. Rankings drop. Traffic dips. But there’s a cost you might not see coming: your brand gets erased from AI answers.

When you change domains, restructure URLs, or move platforms, you’re breaking the thread that connects your content to your entity. AI models like Claude, ChatGPT, and Perplexity don’t just read your words. They build associations between your content, your domain, your links, and who you are as a company. Break those associations, and you stop getting cited.

This isn’t theoretical. We’ve seen clients lose 60% of their AI visibility in the three months after a migration because they treated it like a traditional SEO project. They got the 301 redirects right for Google, but they didn’t preserve their entity signals. AI models couldn’t figure out that the new site was the same company.

Here’s what actually needs to happen to keep your AI presence intact during a migration.

Why Site Migrations Hurt AI Visibility

Traditional search engines like Google understand redirects. A 301 redirect tells Google: “This page moved. Transfer all the authority.” After a few weeks, Google reconnects the dots.

AI models work differently. They’re trained on crawled content up to a knowledge cutoff date. When they respond to queries, they’re referencing that training data. Redirects mean nothing to a model’s training data. What matters is that the entity associations hold up.

Think of it this way. Perplexity reads a Forbes article about your company published three months ago. That article links to your site and mentions your domain. The model encodes that information: “Company X is at domain-name.com. Forbes says they do Y. Their LinkedIn has Z employees.”

Now you migrate to new-domain.com with 301 redirects in place. Google will eventually understand the migration. But Perplexity’s training data still says your company lives at domain-name.com. The model has no way to know you moved. When someone asks about your expertise, the model might cite your old domain, which now 404s in the user’s browser. Or it might not cite you at all, because the entity signal got too weak.

The problem gets worse if you also restructured URLs. Maybe your old site was example.com/services/seo and the new one is example.com/services/search-engine-optimization. Even with redirects, you’ve made it harder for AI models to associate the new content with the old authority signals.

The Pre-Migration Checklist

Start preparing at least 6 weeks before you cut over to the new site. Most teams don’t do this, which is why they struggle.

1. Audit your entity signals. Before you break anything, understand what you’re breaking. Use a tool to crawl your current site and identify every page that has external links pointing to it. These are your high-signal pages. They carry authority. When you change their URLs, you need to be deliberate about it.

Look specifically for:

Create a spreadsheet: old URL, new URL, backlinks, referring domains. You’re building a map of what matters.

2. Preserve your domain if possible. The single biggest factor in maintaining AI visibility is keeping your domain the same. If you’re migrating platforms but staying on the same domain, you’re already 70% of the way there.

If you must change domains (a rebranding, acquisition, or technical necessity), plan for a longer recovery window. You’ll be working harder to maintain entity signals.

3. Document your entity. Before the migration, write down everything you want AI models to know about your company:

This goes into your structured data and your About page. AI models use this to build a mental model of who you are. Make it clear, consistent, and complete.

Implementation: The Technical Side

Set up 301 redirects for every URL with external links. This is non-negotiable. Every page you identified in the audit that has inbound links needs a 301 redirect to its new location. Test each one. A redirect that returns a 302 or 307 instead will hurt you more than no redirect at all.

Why? Because search engines and AI models can follow 301s immediately. They can’t be confident about temporary redirects. They’ll treat your old content as still living on the old domain.

If your entire URL structure changed, redirect consistently. Don’t map old-site.com/page1 to new-site.com/completely-different-page. Map it to the closest equivalent. If there’s no equivalent, redirect to the homepage or a relevant category page. Just map it somewhere sensible.

Update your Organization schema immediately. Before you go live, update your schema markup to include both your old and new URLs. Use the sameAs property:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Your Company",
  "url": "https://new-domain.com",
  "sameAs": "https://old-domain.com",
  "address": {...},
  "contactPoint": {...}
}

This tells crawlers that both URLs represent the same entity. It’s a breadcrumb for AI models during the transition.

If you’re keeping the same domain but restructuring URLs, update all Article and BlogPost schemas to use the new URLs immediately.

Submit your migration to Google Search Console. Google has a dedicated tool for this. Use it. It tells Google: “Hey, we’re moving. Here’s the map.” Google will process it faster than waiting for redirects alone.

Rebuild your internal linking. Update internal links to point to the new URLs, not the old ones. Yes, redirects work. But every redirect adds a hop. Links that point to the new URLs directly are cleaner and signal authority faster.

The Post-Migration Phase

1. Audit AI visibility in the first two weeks. Use a tool to check whether AI models are still citing you. Search for your branded keywords in ChatGPT, Claude, Perplexity, and any other models your customers use. Are you still in the responses? Are the links broken?

If you’re getting cited but with broken links, that’s fixable but not ideal. If you’re not getting cited at all, you need to act faster.

2. Create new high-quality content that links to your established pages. AI models care about topical authority. If you migrate without adding new content, you look stagnant. Create 3 to 5 pieces of new content in the first month that link to your core pages. This signals to AI models that your site is active and that your core content is still central to your expertise.

You don’t need these pieces to rank. You need them to exist and to build internal links that strengthen your entity.

3. Rebuild your citation profile. If you changed domains, reach out to press sites, directories, and industry publications that mention you. Ask them to update their reference to your new URL. A single updated link from a high-authority source is worth 100 directory listings.

Focus on the top 10 sites that link to you. Get those updated in the first month. The rest can follow over time.

4. Monitor redirects for degradation. Set up monitoring to catch broken redirects in the first 90 days. You’ll find mistakes you didn’t catch in testing. A page you thought had no backlinks actually does. Fix these redirects immediately. Each one that breaks costs you authority.

Why This Matters for Answer Engine Optimization

AI search is growing. ChatGPT, Claude, Perplexity, and Google’s SGE are becoming primary discovery channels. Being cited by these models drives brand awareness, trust, and traffic.

Site migrations are one of the few events where you can lose that visibility overnight. A poorly planned migration can erase months of authority building. You go from being cited consistently to barely appearing at all.

The recovery is slower than traditional SEO because you’re fighting against training data. As new versions of models are trained on crawled content, they learn that you moved. But that can take months or longer depending on the model’s update cycle.

Prevention is worth 10 times the effort of recovery. Spend the time now to do this right. Your AI visibility on the other side will thank you.

Key Takeaways

A migration done right is invisible. Your visitors don’t notice. Your rankings stabilize. And most importantly, AI models continue citing you because your entity signal never breaks.

If you need deeper guidance on how AI models evaluate authority or want to understand how to build entity signals that survive migrations, check out our guide to building an entity graph or our AI visibility audit framework.

For technical questions about schema and structured data, see our complete guide to schema markup for AI search. And if you want to understand how AI models rank answers, our breakdown of AI search ranking factors for 2026 covers everything you need to know.