GEO Consultant, New Zealand

What makes a New Zealand page worth citing

AI tools that search the web tend to pick passages that answer the question cleanly and carry some sign of authority. For a New Zealand business, the strongest signal is usually specificity. A generic guide to choosing a freight forwarder or a payroll provider already exists in many overseas versions. A page that explains how the choice works for a New Zealand importer, with local terms, agencies and trade-offs named accurately, offers something those versions lack.

I review each priority page against a short set of questions:

• Does it state something your firm knows first-hand, rather than repeating what every competitor says?
• Are New Zealand terms used correctly where they apply, such as GST-inclusive pricing, building consents or the Companies Office?
• Would a reader in Christchurch or Tauranga recognize the examples as local?
• Is there a named author or reviewer with relevant experience?

Pages that fail every question are rarely worth rewriting line by line. Merging several weak pages into one stronger source, or retiring them, is often the better call. The pages that pass get restructured so the useful part is easy to find and quote, with a clear heading, a direct opening statement and supporting detail underneath.

Turning knowledge you already hold into citable evidence

Original material is what separates a source from a summary. Plenty of New Zealand firms already hold it without noticing: patterns in quoting and job records, faults a service team sees again and again, questions that come up during export documentation, or the steps a project follows from first inquiry to handover. Published carefully, this gives an AI tool something it cannot find on a competitor's site.

The work is mostly editorial judgment. With your team, I decide:

• which observations can be shared without exposing customer details or commercial terms
• whether a pattern is solid enough to publish, or should stay qualitative
• how to describe the method behind any figure you choose to state, so readers can judge it
• who signs off each page and how often it is checked

Anything drawn from customer records should be aggregated and reviewed against the Privacy Act with whoever handles privacy in your business; that interpretation stays with them or your legal advisor.

Comparison and method pages are the other strong format. A table setting out the options buyers weigh, or a step-by-step account of how you deliver, gives a generated answer a structure to borrow. My generative engine optimization page explains the general method; here the focus is the New Zealand evidence behind it.

The outside sources AI answers lean on in New Zealand

Generated answers rarely rely on your website alone. For New Zealand topics they often draw on a familiar mix: government guidance pages, public registers, national and regional news outlets, industry associations and trade bodies, sector directories, and review or comparison sites. Which of these matter depends on your category, so I map them from the citations that actually appear in your prompt runs rather than from a generic list.

For each recurring source, the map records:

• whether your business appears at all, and with what description
• whether services, regions covered or trading name are out of date
• whether you have a legitimate route to a correction or an entry, such as association membership or an editor's contact
• whether an Australian parent, sister company or distributor is cited in your place

I do not buy placements or write disguised advertorials. The aim is accurate, earned presence: an association profile that matches your site, a directory entry listing current services, or commentary offered to a regional publication when you have something genuinely useful to say. Matching official records to your website belongs to my New Zealand AI search optimization work; here the interest is in the sources that shape the substance of an answer.

English, te reo Māori and other language versions that agree

Some New Zealand organizations publish in te reo Māori as well as English, and tourism, education and export businesses often add Chinese, Japanese or Korean pages for overseas visitors. Each version can become a source in its own right. When they drift apart, an AI tool may quote an outdated policy from one language and the current one from another, and the answer becomes unreliable.

I treat this as a parity problem. A source register lists every fact that must match across versions: services, regions served, product names, booking arrangements, policies and contact routes. Each language version is checked against it whenever the English master changes.

The engagement runs in English. Te reo Māori wording comes from fluent colleagues of yours or a translator you trust, with care over place names and macrons. Any Asian-language versions likewise need native speakers rather than machine output. My role is the structure: making sure each version is complete, carries the right language attributes, and is updated on the same cycle as the English.

Not every business needs more than English. If your buyers search and ask in English, extra versions add maintenance without adding citations, and I help you decide that from evidence rather than habit.

Keeping facts fresh, logging citations and knowing the limits

Generated answers age badly when the sources behind them change. Rules, product ranges, service areas and prices move, and a page that was correct when an AI tool last read it can keep being quoted after it stops being true. Each citable page gets an owner, a visible review date and a reminder in a shared calendar, so updates happen before a wrong answer spreads.

Measurement is a dated log, not a dashboard of promises. The same New Zealand buyer prompts are rerun on a schedule across the tools your buyers use, and each run records:

• sources cited and whether any are yours
• how your business is described, and whether that description is accurate
• which overseas or competitor pages are quoted instead

Answers vary between sessions, so I read trends across several runs rather than single results. I also add a discovery question to inquiry forms and a matching field in your CRM, because AI tools do not always pass referral data.

GEO cannot force any tool to cite you, promise a mention or outweigh a poor reputation. It improves the odds by making your pages worth quoting. For the question-and-answer craft alongside it, see my AEO consultant page for New Zealand, or get in touch.