FAQ

Does a corporate website need multiple languages? How do you keep it manageable?

Your website needs a complete version in every language your target market uses to ask AI — engines match citations to the language of the question, so a Chinese-only site is structurally disadvantaged in English answers. Three management rules keep it sane: one page tree, one record, language fields maintained in pairs (two separate trees always drift apart); hreflang annotations so engines know which version serves which market; and a fact pass over every translation, because numbers, dates and credentials must match item for item.

Figure 01 · The hreflang trio: every language version carries the full set

<link rel="alternate" hreflang="zh" href="https://www.example.com/zh/product/" />
<link rel="alternate" hreflang="en" href="https://www.example.com/en/product/" />
<link rel="alternate" hreflang="x-default" href="https://www.example.com/en/product/" />
  1. Self-referentialThe Chinese page must also include the hreflang="zh" line pointing to itself.
  2. CounterpartPoints to the English version of the same content.
  3. x-defaultThe default version (here, the English page), served to users matching no listed language.
  4. Return-link ruleThe English page must carry the same full set and point back to the Chinese page — one-way annotations are ignored by search engines.
Standard hreflang trio (example domain example.com). Specification: Google Search Central: Localized versions.

Why language versions directly affect AI visibility

When users ask ChatGPT in English, engines prefer English sources; when they ask in Chinese, engines do not cite Chinese sources only — but a well-formed Chinese version strengthens entity consistency in Chinese-language contexts. The typical pattern for companies going global: a thin, rarely updated English showcase site and a rich Chinese main site — AI sees exactly the thinnest version of the company. Language versions are not a translation problem; they are a fact-supply problem. AI can only cite the language versions it can read — what it cannot read does not exist.

The reverse is also true: complete language versions that agree are themselves an entity signal. Same company, same facts, two languages: engines cite with more confidence; disagree, and they would rather cite nothing.

Three structures; only the first is recommended

One tree, paired fields (recommended)

One record per page, both languages stored side by side, published, updated and retired together. This site is built that way. Change one fact and both versions change together; retire a page and both go together. AI reads the same facts in either language.

Two independent trees

A separate site per language. Flexible short-term, guaranteed to drift — one side gets edited, the other does not, and AI reads two conflicting versions of your facts and has to guess. The cost is not in building them; it is in doing every later update twice and checking it twice.

Live machine-translation plugins

Cheapest, but translation quality, URL structure and structured data are all outside your control — a negative for AI citation. Machine-translated pages have no stable URL correspondence, and entity descriptions drift with the translation output — engines struggle to treat them as trustworthy sources.

How to get hreflang right

Hreflang tells search engines where the other language versions of a page live. Every language version must carry the full set: its own language, the counterpart language, and x-default (the default version). Two rules matter most: every page must include an hreflang link pointing to itself (self-referential), and language versions must point back at each other (return links) — the Chinese page points to the English page, and the English page must point back. A one-way annotation is ignored by search engines. The figure on this page shows the standard three-link pattern.

Beyond hreflang, both language URLs belong in the sitemap, and JSON-LD entity descriptions must agree across languages. Google's official guidance on multilingual sites is listed under Sources.

Three common mistakes

Mistake 1: unequal content across language versions. The Chinese site carries full product specs; the English site carries slogans. Overseas buyers ask AI in English, and the engine can only read the English surface — the company's most persuasive facts effectively do not exist. Fix: core pages complete in both languages, specs and cases in item-by-item correspondence.

Mistake 2: hreflang without return links. The Chinese page annotates the English version, but the English page never points back. Search engines treat one-way annotations as unreliable, so the correspondence is effectively never established. Fix: check return links appear in pairs at every publish.

Mistake 3: chaotic URL structure. The Chinese and English versions of the same content sit at unrelated URLs (say, Chinese at /product/ and English at /en-goods/) — hard to maintain, and hard for search engines and AI to confirm they are the same page. Fix: design URLs as language pairs (e.g. /zh/product/ and /en/product/), one address per page, stable for life.

Minimum viable setup

  • Core pages (home, solutions, cases, contact) complete in both languages;
  • The full hreflang set (self, counterpart, x-default);
  • Both language URLs in the sitemap;
  • JSON-LD entity descriptions consistent across languages.

News can publish in one language only — just accept it then serves only that language's market. For scale: in our measurement of 30 global-facing websites of Chinese exporting companies (Negency Flagship Report 2026, Issue 1, sampled on a single day, 30 October 2026), 17 had no JSON-LD at all and 13 had a missing or non-unique H1 — entity facts are often missing even in one language, and consistency across languages only raises the bar.

How to verify the language versions are right

Three checks: spot-check that hreflang return links appear in pairs; ask AI the same questions in both languages and compare whether the company facts agree; verify core pages correspond item by item across languages. Done right, AI's answers in both languages should describe the same company. For the rest of the website overhaul, in crawl, structure, facts and trust order, see What should a corporate website change for the AI search era?

See where your brand stands in AI search

Start with the free AI visibility audit: your scores for crawler reachability, machine readability and entity structure in minutes. For the full competitor comparison and fix list, see the AI search visibility diagnostic package.

Related questions

How is GEO different from SEO?

SEO optimises a page's rank in a list of links, aiming for clicks. GEO (Generative Engine Optimization) optimises how discoverable, understandable, trustworthy and citable a company is inside AI-generated answers — the goal is to be in the answer itself. They share technical foundations (crawlability, structured data, content quality), but GEO additionally requires governed entity facts, a machine-readable surface (llms.txt, Markdown mirrors) and a third-party mention ecosystem. A site that ranks well is not automatically visible to AI: engines cite the companies whose facts they can fetch, parse and verify. The concrete differences sit in four dimensions — goal, object, content structure and evaluation — compared item by item in the body below.

How can Chinese companies going global build AI search visibility?

For Chinese companies going global, the visibility gap is rarely content volume — it is the machine-readable surface. Our audit of 30 Chinese outbound companies found machine-readability the weakest category (69.2/100 average): 17 had no JSON-LD on the homepage, 15 had a missing or thin sitemap, 13 had a missing or duplicated H1. The path has four steps: audit bot reachability and machine-readability first; govern entity facts so Chinese and English statements agree; publish question-shaped content built on real overseas buyer questions; then earn third-party mentions (media, communities, partners) and verify with periodic resampling against a fixed question set.

What should a corporate website change for the AI search era?

Five changes. (1) Bot reachability: robots.txt explicitly allows AI crawlers and live fetches are not stopped by firewalls or challenge pages. (2) A machine-readable surface: an llms.txt fact file (optional aid), a complete sitemap, Markdown versions of core pages (optional). (3) Entity structure: JSON-LD naming entity types such as Organization, clean title/H1 discipline, hreflang pairing on bilingual sites. (4) Question-shaped content: pages organised around buyer questions, short answers first, FAQ schema attached. (5) Measurability: analytics that separate human and bot traffic and attribute AI-referred inquiries. Negency's own site was rebuilt to exactly this standard, and the public audit tool scores against it.

What is llms.txt, and should every company website have one?

llms.txt is a plain-text brief at the site root written for AI systems: who the company is, what it does, where the core pages live, and the canonical facts and definitions. It does not by itself raise citation rates — in Ahrefs' May 2026 study of 137,210 domains, 97% of published llms.txt files received zero requests that month, so its direct effect is near zero — but it is the cheapest fact-governance move available: one file anchors a single version of the truth that crawlers can rely on. Verdict: worth adding, useless alone; it works as part of entity-fact governance, structured data and question-shaped content. Negency's llms.txt is a full fact bible: company definition, offering boundaries, sourced cases and citation rules.

What should an enterprise prepare before deploying AI agents?

Four preparations, and we advise against starting without any of them. (1) Scenario selection: pilot on a workflow that is frequent, comparatively rule-clear and verifiable — not the most glamorous one. (2) Knowledge and data foundations: govern the knowledge base first (deduplication, versioning, permissions); a dirty knowledge base only scales errors. (3) Governance design: permission boundaries, human approval points and audit logs decided before launch, not after. (4) A measurement baseline: record the manual process's time and error rates before the pilot, or value can never be proven afterwards. Negency's method is a small controlled pilot, verified before scaling, with governance as a design input rather than a patch.