Has AI Killed International SEO?

Vicky Wilkes

International SEO Strategist

10th November 2026

Share
this article

Any SEO professional using LinkedIn might have got the AI doom and gloom feeling at one point. SEOs and AI have a complicated relationship: while some are scared artificial intelligence might make us redundant, others are excited about the new opportunities that the AI era is bringing (the buzzword here is: at scale).

As an International SEO strategist, I naturally wonder what these changes within our industry mean for my niche. If I trust the strongly opinionated posts on LinkedIn, one could easily believe that the practice of International SEO has died – killed by a deadly combination of the rise of machine translation tools and the ability of LLMs like ChatGPT to translate content within their own system.

Because, let’s face it: AI is not only here to stay (surprise!), but it already is heavily incorporated into our day-to-day SEO workflows. According to Search Engine Land, 72% of SEOs say AI content performs as well as or better than human content and according to the Guardian, even 37% of professional translators use generative AI to support their work. Whether we want it or not, the industry is changing at a rapid pace.

To put it briefly: generative content creation, automated translation tools and in-system translations by ChatGPT and other LLMs are a scary outlook for us International SEOs. So: do we really still need International SEO in the era of AI?

Shaking up traditional workflows

The gold standard when it came to international SEO in the past was always ‘localisation of translation’ and I can’t even count the times that sentence has slipped my lips in various meetings with clients and colleagues.

Google’s John Muller has stated in the past that low-quality translations can harm your SEO efforts, and content that is not only translated well grammatically but truly adapted to your local target audience is the best practice.

However, the current state of play is that more and more businesses are using AI for content translation purposes. Sonix AI state that 39% of marketers are using machine translations, and 83% of those express confidence in the quality of those translations.

We are seeing a shift from manual to AI-assisted workflows within International SEO. From keyword research to plug-ins that can translate full websites at a scale, AI tools are not only cost-efficient, but speedy. And the quality of translations is rapidly improving. This means: our day-to-day processes are changing at a rapid pace.

AI translations – opportunity or false shortcut?

Let’s start with the positive: when it comes to translations, AI systems are not only cost-effective, but also speedy. LLMs can produce human-like translations without requiring any training. While producing content from scratch may take several time-consuming rounds of learning and feedback such as strategic prompt engineering, machine translations can be generated in little to no time. I truly believe that it helps to cut down language barriers and makes content more accessible to non-native speakers.

Outside of word-for-word translations, AI systems fall short. Nuance, cultural relevance and intent alignment are areas that they can’t easily create. And these are, when we look at revenue streams, particularly important. 56% of customers say they feel more loyal to brands that ‘get them’ and show a deep understanding of their priorities and preferences. This is something that I have seen first hand when localising product pages for a big E-commerce client in Germany: We updated the copy to be highly culturally relevant and engaging for the German audience. As a result of that, CVR rose by 92% within 12 months.

Additionally, machine translations can cause serious problems for businesses. Products and services might be regulated by different rules, laws and authorities across multiple markets. In particular, highly regulated industries like pharmaceuticals, financial services, or gambling have to adapt to a vast set of different ground rules. Examples can be as simple as this: while you can market beer to people as young as 16 in Germany, the legal drinking age in the US is 21.

Not taking these things into consideration can not only seriously harm your traditional organic performance through damaging trust and EEAT signals, it can also expose your business to serious legal risks. And that does not take into consideration the chances of replicating a mistake in the origin-copy automatically across multiple market pages.

The echochamber of AI content

Instead of relying on word-for-word translations, you may now come to the conclusion that you could hand over the content writing process to AI, separately for each market.

Let’s consider how AI writes content: by referencing the information that is already available on the World Wide Web and giving you the output that it considers to be most likely relevant for your brand. What would this mean for us if we put ourselves into the shoes of a business that is trying to break into a new market? A market where our product or service may not be available yet, or our offering may not even be culturally engrained with the target audience? Exactly! There is not a lot of information out there to take from.

Less information naturally increases the risk for hallucinations or mistakes, produced by AI. And when we look at the bigger picture, we see an even clearer issue with using AI for international content production: by only building on existing structures, the opportunity for creating something truly new, groundbreaking and resonating with the local target audience is being missed.

Breaking down international SEO structures

The limitations of AI content don’t just affect your workflows — they have consequences for how you’re found in the first place. LLMs become increasingly engrained into the search channel, and Google has rolled out AIOs across 21% of all keywords and is expanding outside English-speaking markets. This, obviously, has a massive impact on International SEO.

While we know that local SERPs clearly favour localised content, LLMs have the ability to translate content within their own systems. Technical structures are losing grip, as LLMs don’t read websites like traditional crawlers do.

Classic technical international SEO tools like hreflang, ccTLDs or server signals become less relevant, as LLMs are working by converting semantic concepts into numerical vectors.

This is something we can observe while using ChatGPT in real time: ChatGPTs answer to my German question “What are the seven seas, and what should you know about them?” only states sources in English language:

A study done by Peec AI demonstrates this very clearly: when users research in German, Polish, or Spanish, half of the research that ChatGPT carries out happens in English. This can have a devastating impact on businesses, as it can lead to companies being completely invisible in ChatGPT answers, even when they dominate local search results. Heading towards a search-world where ChatGPT is more and more engrained into the purchase research process, this might put local businesses far behind bigger corporations.

This leads me to two main conclusions:

Organic SERPs are as important as ever, particularly outside English-speaking markets. Businesses need to combat potential biases that ChatGPT might have towards non-English-speaking sources
The concept of one URL for one language-specific target audience and market does not apply when we want to generate visibility within LLMs.

International visibility within LLMs

As said above, LLMs can translate content within their system and rely on semantic and contextual relevance and not on classic ranking signals. They reward entities that are unambiguous and well-documented across the web, rather than translated into different languages.

However, we need to keep in mind that the actual ‘solution’ that LLMs provide for users within the search channel is carrying out research on their behalf. Instead of providing them with a bunch of relevant sources, like traditional SERPs, they display all the information a user might be looking for in a hyper-relevant text result.

The keyword here is: hyper-personalisation. LLMs want to provide answers that are highly relevant and personalised to the target audience.

Let’s go back to the example of a business trying to expand into a new market. In our example, this is a price comparison website for everyday products. A tool like that is heavily engrained into German culture, people use it every day to purchase products like smartphones, gadgets, or appliances. Meanwhile, in the UK, comparison tools are not as popular and mainly used for financial services.

When I ask ChatGPT in German “How can I find the best deal for a Dyson Airwrap?”, price comparison tools are not only heavily cited, but 3 providers are immediately listed:

When I ask in the UK, this looks completely different. Price comparison websites are not even mentioned, but instead, voucher providers and traditional retailers dominate the answer:

While comparison tools exist in the UK, and voucher pages exist in Germany and content is available across both languages. But to put it simply: ChatGPT does not consider comparison tools to be relevant in the UK, and voucher pages to be relevant in Germany. Businesses operating in these industries remain completely invisible within LLM replies.

This shows us that new factors become more important when it comes to generating visibility within LLMs. Entity clarity is the most impactful tool we have. Hyper-personalised content that can be pulled into contextually relevant answers is the key to LLM visibility – and local content feeds this.

Instead of thinking: how can I localise my content to reach local audiences and be more visible within LLM, we need to move to full transcreation processes: we need to ask ourselves which value can we add to make our content hyper relevant and personalised and how can we engrain our offering into the local culture.

A high-quality translation with slight cultural nuance won’t bring the desired efforts, we have to think bigger.

Not all AI systems are the same

Additionally, we can’t assume that all AI systems operate the same way. Something I have discovered while working with clients on the German market and that has been observed by other International SEO experts too is that Google AIOs focus more heavily on local and native-language or regional authority signals.

Another aspect that is important to consider: the way people search impacts the output of the AIO. When looking at phone insurances, for example, Germans search for the short query “handyversicherung sinnvoll”, which translates to “phone insurance useful”. The AIO is very direct: no, phone insurance is generally not useful, apart from certain cases. The conclusion is clear: most people don’t need it. Cited sources are authoritative, like the Consumer Advice Bureau and financial advice websites.

In the UK, a more conversational search term is more popular: “is phone insurance worth it?”. The AIO overview here is more conversational too: the usefulness of a phone contract depends on the user, there are a lot of pros and cons to consider – the answer is not as straightforward and the AIO links to comparison websites where readers can look at pro and contra lists in full detail.

This means: One localisation approach won’t fit all. We need to truly understand which AI tools our audiences are using, how they search in them, and what kind of sources different LLMs consider to be relevant

International SEO in the era of AI – not dead, but different

International SEO isn’t going anywhere. But we need to think it differently and move away from the “One URL, one language” approach, supported by technical tools.

We need to ask ourselves: instead of translating or localising my content, what can I provide for my local target audiences to make it truly relevant and hyper-personalised to them? AI has raised the stakes: What’s emerging is a more sophisticated discipline that blends technical SEO, cultural intelligence, and AI fluency.

The era of translated content pages is over.

Key takeaways

AI translations have real limits: Machine translations are fast, cost-effective and improving in quality, but they struggle with cultural relevance, intent alignment and trust-building. They can be a great start when you want to translate a website from scratch, but to be successful in the long term, you will need to consider impactful transcreation efforts
Technical SEO signals are losing grip: Classic tools like hreflang are still meaningful when it comes to organic clicks, but the “one URL, one language” approach doesn’t apply to LLMs. You need to consider a more holistic optimisation approach for LLMs to understand your site’s intent.
LLMs carry biases toward English-language sources:. Even when users search in other languages, ChatGPT frequently draws on English content, meaning businesses that dominate local SERPs can even be invisible in LLM-generated answers. If this is the case for you, you need to ensure that your content is hyper-relevant for local audiences.
Entity clarity and hyper-personalisation are the new priorities: LLMs reward brands that are well-documented and unambiguous across the web. Content needs to go beyond localisation into full transcreation, deeply embedding a brand into local culture and context.
Different AI systems behave differently: Google AIOs and ChatGPT weigh signals differently, and search behaviour varies by market, meaning no single localisation strategy works universally.

___________

Further reading