A NZ SaaS founder searches for their own company in ChatGPT. Nothing. They search for the problem their product solves. A competitor - an Australian company, based in Melbourne - gets named. They've never heard of them. The Melbourne company isn't bigger, better-funded, or better at SEO. They just set up their site to be readable by AI. That gap is now a commercial problem, and it's fixable in weeks.
Why NZ B2B is behind on AI visibility
It's not incompetence. NZ B2B marketing teams are lean by necessity - a team of one or two doing the work of five. AI search optimisation hasn't been commercially urgent long enough for it to land on the priority list. And most of the content about AEO comes from US or UK sources, making it easy to assume it's not relevant yet in this market.
It's relevant. Buyers researching B2B software in NZ are using the same AI tools as buyers in San Francisco. When a Wellington CFO asks ChatGPT to recommend a SaaS tool for their finance team, the AI doesn't filter results by geography - it surfaces whoever has built the best AI-readable presence. Right now, that's almost never a NZ company.
"The NZ B2B companies I've audited are, almost without exception, optimised for 2018 Google. That's not a criticism - it reflects where the tooling and guidance have been. But the window to move first on AI visibility is closing."
What "invisible to AI" actually means
Being invisible to AI doesn't mean your site doesn't exist. It means that when an AI engine is constructing an answer to a question in your category, it either can't read your content reliably, doesn't trust it enough to cite it, or doesn't know your company is relevant to the question being asked.
There are three distinct failure modes, and most NZ B2B sites have all three simultaneously.
It can't be read properly
AI crawlers read differently to Google's crawler. They prioritise structured, machine-readable signals: schema markup that declares what your content is, an llms.txt file that gives them a plain-language summary of your site, and clean semantic HTML that makes content hierarchy unambiguous.
Most NZ B2B sites have either no schema or schema that was auto-generated by a plugin and never validated. Almost none have an llms.txt. The result is that AI crawlers land on the site, can technically read the words, but have no reliable signal about what the site is, who it's for, or which pages matter.
It can't be trusted enough to cite
AI engines build authority models. They consider whether your content is consistent, specific, and corroborated across multiple sources. A company whose homepage says one thing, whose LinkedIn says something slightly different, and whose G2 listing says something else gives the AI conflicting signals - and conflicting signals lower trust.
This is the entity consistency problem. It's not about backlinks or domain authority in the traditional SEO sense. It's about whether the AI can construct a coherent, reliable picture of who you are and what you do from the signals available to it.
It's not seen as relevant to the question
Even if an AI can read your site and trusts your content, it still needs to understand that your product or service is the answer to the specific question being asked. That requires your content to be structured around questions - not around keywords.
A page optimised for the keyword "project management software NZ" may rank fine on Google. But if a buyer asks ChatGPT "what's the best project management tool for small NZ engineering firms," the AI needs content that directly addresses that specific question - the use case, the company size, the industry, the geography. Keyword-optimised content rarely does this well.
The NZ-specific opportunity
Here's the read that most NZ founders miss: being a smaller market is actually an advantage right now. The AEO moat is available to whoever moves first, and in most NZ B2B categories, nobody has moved yet. In larger markets - US, UK, Australia - the early movers are already established and compounding.
A NZ SaaS company that implements solid AEO fundamentals in Q2 2026 can own its category's AI citations within 3–6 months. That's not a theoretical outcome - it's what I've seen happen in categories where a single company moved early and built the foundation properly while competitors were still thinking about it.
The international ambition angle
For NZ companies with AU or global expansion on the roadmap, this matters even more. An Australian buyer using AI to research vendors doesn't know or care that you're based in Auckland. If your AEO foundation is solid, you show up. If it isn't, a Sydney competitor who set it up six months ago is the one getting cited.
AEO doesn't care about geography. It cares about signal quality. A NZ company with strong schema, a well-structured llms.txt, direct answer-led content, and consistent entity signals will outperform a larger Australian competitor with none of those things.
The four fixes - in priority order
This isn't a 12-month project. For most NZ B2B sites, the core AEO foundation can be implemented in 4–6 weeks by a focused operator. Here's the priority order, based on impact per hour of work.
Fix 1: Schema markup audit and rebuild
Start here. Run your site through Google's Rich Results Test and the schema.org validator. You will almost certainly find errors, outdated markup, or critical missing types. The minimum viable schema set for a B2B SaaS site is: Organisation, WebSite, WebPage (typed per page), Product or SoftwareApplication, FAQPage on relevant pages, and BreadcrumbList throughout.
Each schema block should be comprehensive - not just the required fields, but the recommended fields that help AI engines understand context. An Organisation schema with a name and URL is barely better than nothing. One with a full description, founding date, employee range, sameAs links to LinkedIn and Crunchbase, and area served is a completely different signal.
// Paste in your site <head> as JSON-LD { "@context": "https://schema.org", "@type": "Organization", "name": "Your Company", "url": "https://yourcompany.co.nz", "description": "One clear sentence: what you do, for whom.", "foundingDate": "2021", "areaServed": ["NZ", "AU"], "sameAs": [ "https://linkedin.com/company/yourcompany", "https://g2.com/products/yourcompany" ] }
Fix 2: Write and publish llms.txt
This is a 30-minute task that almost nobody has done. Create a plain text file at yoursite.co.nz/llms.txt. Write 200–400 words in plain English describing: what your company does, who your ideal customer is, what problems you solve, which pages are most important, and what you'd want an AI to know about you before answering a question about your category.
Write it like you're briefing a smart assistant who has never heard of you. No marketing language. No jargon. Clear, specific, direct. The companies doing this are getting cited more frequently - not because the file is magic, but because it removes ambiguity that was previously costing them citations.
Fix 3: Restructure content around questions
This is the most time-intensive fix but has the longest-lasting impact. Go through your top 10 pages and audit them against this test: does each H2 read like a question a buyer would actually ask? Does each section open with the answer rather than building toward it? Is the specific use case, customer type, and outcome named explicitly?
For most NZ B2B sites, the answer to all three is no. The fix isn't a full rewrite - it's restructuring. Move the answer to the top of each section. Rewrite H2s to be question-shaped. Add a FAQ section to your pricing and product pages with real buyer questions as the schema-marked questions.
Fix 4: Consolidate entity signals
Audit how your company is described across every platform where it appears: your website, LinkedIn company page, Crunchbase, G2 or Capterra, any media mentions, any partner listings. The description of what you do, who you serve, and what you're called should be identical in substance across all of them.
This sounds basic. In practice, most companies have four or five different descriptions of themselves that evolved organically over time, and none of them are optimised for the question an AI is trying to answer. Standardising these signals is low effort, high impact, and permanent.
What 500% organic growth actually looks like
The 500% organic B2B traffic growth I've achieved for a client in 6 months didn't come from one big unlock. It came from implementing all four of the above fixes in sequence, then compounding the gains with consistent answer-led content production - one to two pieces per month, each structured around a specific buyer question in the category.
Week 1 to 5: foundation work - schema, llms.txt, entity consolidation, content restructure. Traffic flat or slightly down as crawlers re-index. Week 6 to 12: first citations start appearing in AI tools. Organic search starts climbing as the restructured content begins ranking for long-tail queries. Week 13 to 24: compounding. AI citations drive branded searches. Branded searches drive more citations. Traffic reaches 3.5x baseline and is still climbing.
The timeline isn't guaranteed - category competitiveness, existing domain authority, and content volume all affect it. But the direction is consistent across every implementation I've done: foundation first, then compound.
Where to start this week
- Run the schema audit: Google Rich Results Test + schema.org validator. Note every error and missing type. This is your punch list.
- Write your llms.txt: open a doc, write 200 words in plain English about what you do and who for. Publish it at /llms.txt. Done.
- Pick your top 5 pages: homepage, product, pricing, and your two best-performing blog posts. Restructure each one with question-led H2s and answer-first paragraphs.
- Audit your LinkedIn company description: compare it word-for-word to your homepage. Standardise the substance. Do the same for G2 and Crunchbase if you're listed.
- Add FAQPage schema to your pricing page: take the three questions you get asked most on sales calls. Put them on the page with direct answers. Mark them up with FAQPage schema.
That's 2–3 days of focused work. It won't get you to 500% in week one. But it will get your site off the invisible list - and in most NZ B2B categories right now, that's enough to be the only one the AI finds.
NZ B2B sites are almost universally invisible to AI engines. That's a problem and an opportunity at the same time.
Schema, llms.txt, question-led content, and entity consolidation are the four fixes. Most NZ competitors haven't done any of them. The window to own your category's AI citations is open right now - and it won't stay open indefinitely.
