Your German website may contain years of useful knowledge: detailed product explanations, practical guides and answers to questions your customers regularly ask. Its English version might cover only the company introduction and main products.
Now imagine an English-speaking customer entering a question into a search engine and receiving an AI-generated summary that draws on your German content. If the search engine can present that information in English, does the original page still need translating?
It’s a reasonable question for your multilingual content strategy. However, a search engine’s ability to work across languages doesn’t tell you whether it will select your content, how it will represent it or what the customer will find when they visit your website.
The practical issue is which parts of your knowledge English-speaking customers should be able to access directly.
Can AI Search Use German Content to Answer English Questions?
AI systems can draw on information in one language and produce an answer in another. However, two separate decisions affect what the reader receives:
- Source selection: which documents the system finds and uses.
- Answer generation: how it selects, combines and expresses information from those documents.
A fluent English answer may draw on German material. Conversely, an answer may reproduce awkward, literal translations of German expressions. The language of the response alone doesn’t establish how thoroughly the source has been understood or represented.
This creates possibilities for cross-language discovery, but it also leaves uncertainty. A system may be capable of using your German guide without choosing it for a particular English question.
What Does the Research Show?
Research suggests that an AI system may be able to understand several languages while still favouring sources written in the language of the question.
The study Faux Polyglot, published at NAACL in 2025, examined systems that retrieve documents and use them to generate answers across languages. It found a preference for information in the same language as the question, both when retrieving documents and when generating responses.
A separate study, Linguistic Nepotism, examined how six language models selected citations from sources across eight languages. For English questions, the models preferentially cited English sources and sometimes favoured language over document relevance.
These were controlled research settings. They don’t establish how every commercial AI search tool behaves, or prove that translating a page will increase its citations.
They do support a cautious conclusion: the ability to process several languages doesn’t mean that information in those languages receives equal consideration.
Does That Reduce the Need for English Content?
An AI-generated answer may help someone discover information from your German website. However, that answer is only one possible part of their decision.
A prospective customer may want to inspect the source, compare products in more detail or send a colleague a useful explanation. If they follow a citation and reach a German-only page, they may find it difficult to continue.
Some readers may be comfortable relying on the information in the AI-generated answer. Others may prefer a source they can read directly, particularly when comparing detailed requirements or checking a condition before ordering.
Publishing an English version gives them a page whose wording your business has reviewed and approved. You cannot control how another system summarises it, but you can provide a clear reference for anyone who wants to explore further.
Language also remains relevant to conventional search. Google’s guidance on multilingual websites states that Search tries to find pages matching the searcher’s language. That is broader search guidance, rather than a promise about AI citations, but it remains relevant when planning how customers can find your content.
None of this means that every German page needs an English counterpart. It means that translation priorities should account for what happens both during discovery and afterwards.
Which Content Deserves a Place in Your Multilingual Content Strategy?
A useful starting point is to ask which customer decisions your English website currently supports.
Your main sales pages may explain what you provide. Supporting content can help readers establish whether it suits their circumstances and understand the reasons for choosing one option over another.
Core Product and Service Information
Check whether English-speaking customers can understand the relevant range, its intended uses and the requirements that affect suitability.
A product name and short description may be enough for an initial overview. They may be insufficient when a customer needs to compare preparation methods, compatibility or service arrangements.
The appropriate depth depends on what the reader needs to decide.
Guides That Help Customers Assess Their Options
A German buying guide may contain explanations that your English product pages lack.
For example, it might explain differences between preparation methods, outline what to check before selecting a product or describe situations in which one option is more suitable than another.
If those questions are also relevant to English-speaking customers, the guide deserves consideration. Prioritise the guide according to how useful it would be to those customers. Having a detailed German version ready to translate is helpful but doesn’t necessarily make translating it a higher priority than addressing other gaps in your English content.
Specialist Knowledge With a Clear Customer Purpose
Some of your most useful material may sit outside the main sales pages: an explanation written by a product specialist, a practical comparison or a detailed response to a recurring enquiry.
Consider whether this content helps customers understand something consequential. An article that resolves a common source of uncertainty may warrant translation before a general company news item.
For advice on identifying customer questions and using confirmed product information to answer them clearly, see my article exploring how to make website content answer your customers’ questions.
Example: Looking Beyond the Core Product Pages
The following constructed scenario illustrates a translation-planning decision. It is not a test of AI search performance.
Imagine a German supplier of professional skincare products expanding its English website to reach UK beauty therapists who offer facial treatments.
Its current content includes:
| Content | German Version | English Version |
| Company and product-range introduction | Available | Available |
| Individual product pages | Detailed descriptions and specifications | Short descriptions |
| Guide to comparing facial cleansers | Available | Not available |
| Product training dates in Germany | Available | Not available |
The German cleanser guide explains how products in the range differ in texture, removal method and intended use within a treatment. It links to individual product pages for the relevant specifications.
An English-speaking beauty therapist researching cleansers could potentially see an AI-generated search summary drawing on that guide. However, the supplier has no basis for assuming that the system will select it or reproduce the distinctions clearly.
The planning question is therefore: would making this information available directly in English help the customers the supplier wants to reach?
When Translating the Guide Makes Sense
Suppose the products discussed are available in the UK, the guidance applies to those customers and the sales team regularly receives questions about choosing between them.
Those circumstances give the guide a clear purpose. An English version could help beauty therapists compare the range before looking at individual products or contacting the supplier.
A translation would need to reflect the products and information relevant to that market. Any differences in availability or approved claims should be clarified with the business as part of the brief.
When Updating Existing Pages Comes First
However, suppose the English product pages omit the removal methods and intended uses that the guide refers readers to.
Translating the guide alone would mean that customers following its links would still find incomplete product information. Improving those pages may therefore be the first priority, with the guide following as a second stage.
There is also no immediate reason to translate training dates for events delivered only in German and intended for German-speaking attendees, unless English-speaking customers also need that information – for example, because they speak German and could attend.
The resulting plan is selective. It follows the customer’s information needs across the website, rather than treating every untranslated page as an equally important gap.
How Should You Set Translation Priorities?
You can begin with a small review rather than a commitment to translating the entire website.
Identify the Questions That Matter
Choose a product range, service or customer group that is commercially important to your business.
Then identify the questions people need answered before they can assess suitability or take the next step. Customer enquiries, sales conversations and website search data may help you establish these.
Keep the questions specific enough to guide a decision about content.
Check Where the Answers Currently Live
For each question, establish:
- Whether your website answers it at all
- Whether the useful information is available in German, English or both
- Whether it remains current and relevant to the intended English-speaking customers
- Whether readers can reach the related specifications or next steps in English
This can reveal different kinds of work. A German guide may need translation, an English page may need expanding, or outdated source content may need revision before either task begins.
Choose the Work That Resolves the Most Useful Gap
Prioritise content that combines a clear customer need with reliable information and a relevant next step.
An older article that explains a persistent buying question may be more useful than a recent announcement. Equally, adding one missing explanation to an existing English page may be more proportionate than translating a long guide.
Is an Automatically Translated Version Enough?
Making a page available in English is a starting point. The next question is whether English-speaking customers can use it to understand the same distinctions and make an informed decision.
Machine translation may leave overly literal wording that obscures a practical explanation or misrepresents specialist terminology. My article on the risks of literal translation explores why English wording can need more substantial changes.
Generative AI can produce fluent English, but that fluency still needs scrutiny. A draft may omit a condition, blur a product distinction or introduce an explanation the source doesn’t support. My guide to reviewing an AI translation before publication explains how to compare the draft with the German source and assess whether the English serves its intended purpose.
For the skincare supplier, that means checking whether beauty therapists can accurately compare the products using the English content. The method used to produce the draft doesn’t establish whether it is ready to publish.
If you’re commissioning support, agree which pages will be translated, where adaptation is needed and which factual questions your team must resolve. A translator or editor can help assess the material and shape the English; your business supplies or confirms the underlying product and service information.
AI search gives you another reason to examine where your expertise is available across languages. The strongest case for translating a particular page remains concrete: it gives the customers you want to reach information they can use.
If you’d like help reviewing which German content would be most useful in English, get in touch with a little background on your website and intended audience.
