Your New Website Could Make You Invisible to AI. Here’s Why.

Your Website Migration Is an AI Visibility Event. Most Brands Don’t Treat It That Way.

For years, the risk calculation around a website migration was relatively straightforward. Get your redirects wrong, and you’d lose rankings. Forget to update your sitemap, and Googlebot would take longer to find your new pages. Miss a canonical tag, and you’d spend a few weeks mopping up duplicate content issues.

These were, and still are, serious problems, but they were recoverable problems. Rankings could be rebuilt. Crawl coverage could be improved. Traffic, in most cases, came back – eventually.

That risk calculation has changed. Not because the technical requirements of a migration have become more complex, though in many cases they have. It has changed because the way brands are discovered online has fundamentally changed, and a website migration now has consequences that reach far beyond your Google Search Console impressions report.

When a brand migrates its website today, it is not just moving pages. It is potentially disrupting the entire body of evidence that AI systems use to understand, trust, and recommend that brand. And unlike a rankings drop, that kind of disruption is not always visible and not always straightforward to fix.

How AI Understands Your Brand

To understand why migrations carry a new category of risk, it helps to understand how AI search systems form a picture of your brand in the first place.

AI models, whether that’s Google’s AI Overviews, ChatGPT, Claude, Gemini, or Perplexity, do not simply read your website and decide what you do. Whilst they all use slightly different approaches to retrieval, indexing and training they build an understanding of your brand through repeated encounters with consistent information across multiple sources. Your own website is just one input. But so are the third-party articles that reference you, the directories that list you, the case studies that describe your work, the press coverage that quotes your team, and the forum discussions where your brand gets mentioned.

The more consistently all of those sources describe your brand in the same way; the same services, the same specialisms, the same geographic focus, the same proof points, the more confidently an AI system can represent you accurately when a user asks a relevant question.

This is what is sometimes referred to as entity clarity. Your brand is an entity. AI systems build their understanding of that entity from the sum of what they encounter about you across the web, weighted by how credible and consistent those sources are.

A well-executed migration, in this context, preserves that entity clarity, a poorly executed one can fracture it completely.

What a Migration Can Break That You Won’t See in Search Console

The most visible consequence of a bad migration is a traffic drop. Rankings fall, impressions decline, and the organic channel takes a hit that can take months to recover from. This is serious, but it is at least measurable. You can see it, diagnose it, and work to fix it.

What is harder to see and harder to fix is what happens to your brand’s AI visibility when a migration goes wrong.

Broken evidence chains

Many of the third-party sources that AI models trust; news articles, industry publications, directory listings, review platforms etc link to specific pages on your website. When those pages move and redirects aren’t implemented correctly, the chain of evidence that connects your brand to its areas of expertise is broken. The AI model’s system can no longer access the credible sources or references. Over time, that gap gets filled; sometimes with a competitor, sometimes with outdated information, sometimes with nothing at all.

Entity confusion

AI systems build brand understanding through co-occurrence, the repeated proximity of your brand name to specific topics, services, and attributes. If a migration causes your key service pages to temporarily disappear from the index, to be replaced by thin placeholder content, or to lose the structured data that made them interpretable, the co-occurrence signals that AI relies on are disrupted. Your brand may continue to exist in AI training data, but its association with specific topics weakens.

Inconsistent brand description

One of the most common and most underestimated consequences of a migration is that brand descriptions become inconsistent across the web. The new website positions the brand differently to the old one. Old third-party articles describe services that no longer appear on the site. Directory listings reference a strapline that was retired in the redesign. To a human reader, this looks like a brand refreshing its positioning. To an AI model attempting to build a coherent picture of what a brand does and for whom, it looks like contradictory information and AI systems resolve contradictions by defaulting to the sources they trust most, which are rarely the brand’s own website.

Structured data loss

Many migrations, particularly those involving platform changes, result in structured data being stripped, overwritten, or simply not reinstated. Schema markup that told search engines and AI systems exactly what a business does, where it operates, what its products and services are, and what proof exists of its credibility, disappears. The information is still there in the page copy, but it is no longer machine-readable in the same way. This matters more now than it did two years ago, because structured data helps AI systems interpret information more consistently.

The Timing Problem

There is a timing dimension to this that makes the stakes higher than most brands realise. AI models are not updated in real time. They are trained on data up to a certain point, and then they go through a cycle of retraining and updating that varies by model and by platform. Some systems update via retrieval/index refreshes rather than needing full retraining, however if your migration causes a period of confusion; broken pages, inconsistent signals, thin content etc that confusion can become baked into an AI model’s understanding of your brand for a period of time, even after you’ve fixed the underlying technical issues.

A rankings drop from a bad migration is typically recoverable within weeks or months once the technical problems are addressed. A distorted AI understanding of your brand can persist far longer, because you’re not just waiting for a crawler to revisit your pages, you’re waiting for a model to be retrained on corrected information.

This situation, where the damage is faster than the recovery, is one of the strongest arguments for getting a migration right before it goes live, rather than fixing it afterwards.

What a Migration Should Protect

Given all of this, the scope of a well-managed migration needs to be broader than most technical SEO checklists currently reflect. Koozai’s SAFE SEO Migration Method™ already includes both the established requirements; redirect mapping, crawl coverage, sitemap updates, page speed, canonical tags, etc and takes AI visibility seriously by addressing the following:

Structured data continuity

Schema markup on the existing site is noted, carried over to the new site, and reviewed for accuracy against the new content structure. If the migration involves a platform change, structured data should be treated as a key deliverable, not an afterthought.

Entity signal preservation

The pages that carry the most entity value, for example service pages, about pages, location pages, author profiles etc should be identified before the migration and treated with high priority for redirect accuracy, content continuity, and structured data reinstatement.

Third-party references

Identifying the external sources that reference specific URLs on the current site; backlinks, citations, directory listings, press coverage. Where those references point to pages that will move, a plan should exist for ensuring the redirect is clean and, where possible, for updating the reference at source.

Brand description consistency

If the migration involves a repositioning or a change in how the brand describes its services , a parallel exercise should map out every place on the web where the old description appears and a plan should be developed for updating those references over time. This is not a one-day job, but it needs to start at the point of migration.

The Broader Principle

Website migrations have always been high-stakes moments. The difference now is that the stakes extend beyond what we’re used to.

AI-generated answers are increasingly the first point of contact between a brand and a potential customer. They are shaping purchasing decisions, informing B2B research, directing patients to healthcare providers, and guiding travellers toward booking choices, often before a single website has been visited. The brands that appear in those answers, accurately and consistently, have a significant advantage. The brands that don’t; whether because they were never there, or because a migration disrupted the evidence that put them there, are invisible in a way that doesn’t show up on a dashboard.

A website migration isn’t just a technical event with SEO consequences, it’s a moment at which a brand’s entire digital evidence base is at risk. Treating it accordingly, with the right expertise involved from the start, the right scope of work, and the right monitoring afterwards is not optional for any brand that takes its visibility in the next generation of search seriously.

Koozai provides website migration SEO for UK businesses planning a site relaunch and emergency recovery services for brands where a migration has already caused traffic loss. We have published a Website Migration SEO Rescue Guide for brands dealing with post-migration issues, and our work includes migration projects for Designlab and Penson. If your brand is planning a migration or has experienced a traffic drop following a recent relaunch, speak to our team.