Your customers aren’t just Googling anymore. They’re asking ChatGPT, Gemini, Perplexity and Google’s AI Mode for recommendations, and getting one answer instead of ten blue links. If your brand isn’t in that answer, then someone else is being positioned as the solution.
Our AEO services help New Zealand brands get found, cited and recommended by AI. It’s called answer engine optimisation (AEO) or generative engine optimisation (GEO), depending on who you ask. We’ve been doing SEO since the early 2000s, and AI search engine optimisation is the natural next chapter.
What is AEO (and GEO)?
Answer engine optimisation (AEO) is the practice of making your brand, products and content easy for AI-powered search tools to find, understand, trust and recommend. Classic SEO aims for a ranking position. AEO aims to get you mentioned, cited and recommended inside the answer itself.
The industry hasn’t settled on an acronym yet, and we’ve decided not to lose sleep over it. The term “GEO” comes from a 2023 research paper by Princeton and collaborators. “AEO” goes back to the days of featured snippets and voice assistants. It’s all the same, but you can call us an AEO agency.
How AEO works
Ask an AI assistant a question and a surprising amount happens in the few seconds before it replies. Broadly, it does four things, and AEO is about influencing all four.
-
Step 1
It reads the question, and the context around it
AI assistants don’t treat a question in isolation. They factor in earlier messages in the conversation and, increasingly, what they’ve learned about the person asking. ChatGPT can reference past conversations to tailor its responses (OpenAI), and Gemini can draw on your Google Search history if you allow it (Google). Two people asking the same question can get quite different answers, so brands need to be relevant to many versions of a question, not just one keyword.
-
Step 2
It fans out into lots of searches
Rather than running one search, the AI breaks the question into sub-topics and searches for each of them. Google describes AI Mode’s “query fan-out” technique as “breaking down your question into subtopics and issuing a multitude of queries simultaneously” (Google). A question about a Wellington accountant might trigger searches for pricing, reviews, sole-trader services and GST help all at once. Each of those searches is a chance for your brand to show up, or to go missing.
-
Step 3
It works out who’s who
Search results are just text until the AI matches each mention to a real business. It blends what it finds now with what it already learned in training, which can be months or years out of date. If your name is shared with another company, a place or an everyday word, or your details differ from site to site, the AI may merge you with someone else, fill the gaps with old facts, or leave you out because it isn’t sure it’s you. Making sure it gets this right is called entity optimisation.
-
Step 4
It synthesises the consensus into one answer
Finally, the AI reads what comes back, compares the sources and writes a single answer. Claims backed up by several independent sources carry more weight than a brand talking about itself. That’s a big reason a brand’s own website makes up only 5–10% of the sources AI search references (McKinsey). The answer usually recommends a handful of brands and cites a few of the sources it relied on.
We make your brand relevant to the many ways people ask, findable across the searches AI runs behind the scenes, unmistakable when AI works out who’s who, and backed by a consensus that’s accurate, positive and easy for AI to find.
Why AEO matters
The front door to your brand has moved. AI search went from novelty to habit remarkably quickly, and Kiwis are right in the thick of it. Toggle between the global picture and what’s happening here in Aotearoa:
of NZ online shoppers use AI tools at least sometimes when shopping (61% of 18–39s).
IAB Australia & Pureprofile, 2026 (n=877 NZ)of NZ online shoppers use AI summaries for a quick answer before clicking through to a website.
IAB Australia & Pureprofile, 2026of Kiwis surveyed had used an AI chatbot instead of a search engine.
IAB NZ, 2025 (n=250, upper North Island)lower click-through rate for top-ranking pages when a Google AI Overview appears.
Ahrefs, 2026 (300k keywords)higher conversion rate for AI-referred retail traffic than for non-AI traffic.
Adobe, May 2026 (US retail)of the sources AI search references come from a brand’s own website. The rest is what others say.
McKinsey, Oct 2025Before anyone declares SEO dead: Google still sends websites roughly 190x more traffic than ChatGPT does (Ahrefs, 2026). AI isn’t replacing search. It’s reshaping it, which is exactly why AEO and SEO belong in the same plan.
Our AEO services
AEO covers tech, content, PR and analytics, so we treat it as one joined-up programme rather than a bag of tricks. Pick an AEO service to see what’s involved:
Technical accessibility for AI crawlers
Open the door (to the right bots).
If an AI crawler can’t reach or read your content, you simply don’t exist in the answer. None of the major AI crawlers currently render JavaScript (Vercel), and Adobe found around a third of the average US retail product page isn’t machine-readable (Adobe, 2026).
- Audit robots.txt, CDN and firewall rules for AI search and training bots (OAI-SearchBot, GPTBot, PerplexityBot, ClaudeBot, Google-Extended and friends)
- Check what AI crawlers see without JavaScript, and fix it with server-side rendering or pre-rendering
- Log file analysis to see which AI bots visit, how often and what they fetch
- Key facts like prices, locations and specs in plain HTML, not locked inside scripts or images
- Bing Webmaster Tools and IndexNow set-up, because Bing’s index powers Microsoft Copilot
Entity optimisation for AI
Make sure AI knows it’s you.
Before AI can recommend you, it has to recognise you, tell you apart from anything with a similar name, and know which categories you belong in. Language models build that picture from two places: what they absorbed in training, and what they find in live searches. When those disagree, or your details vary from source to source, you see the classic AI mistakes.
- An AI entity audit: what ChatGPT, Gemini, Perplexity and AI Overviews say you are, do and sell, checked against the truth
- Disambiguation for brands that share a name with a place, an everyday word or another business
- Category association, so AI links you to the problems you solve, not just your name
- Strengthening the reference sources models lean on: Wikidata, Google Business Profile, LinkedIn, industry directories, and Wikipedia.
- Organization, Product and Service schema with
sameAslinks as a supporting signal (it’s part of our SEO brand and entity work too)
Prompt & query research
Learn how Kiwis actually ask.
People don’t type “accountant Wellington” into ChatGPT. They say “I’m a sole trader in Wellington who hates paperwork, who should I talk to?” Behind the scenes, the AI then runs its own searches. One study found ChatGPT searched the web on 31% of prompts, averaging 2.17 searches each, and leaned heavily on reviews, comparisons and local intent (Nectiv via Search Engine Land, 2025).
- Build prompt libraries by persona, buying stage and intent, in real Kiwi language
- Map the “query fan-out”: the hidden searches AI runs to build its answers
- Identify which sources get cited for your topics, and which of them mention competitors but not you
- Turn your existing keyword and Search Console data into conversational prompt sets
Content strategy (and delivery)
Be the best answer, not just an answer.
AI engines favour content that’s clear, specific and verifiable. Research shows that adding citations, quotations and statistics boosted visibility in generative engine responses by up to 40% (Aggarwal et al., KDD 2024). We plan the content strategy, and then (if you want us to) we can actually write the content.
- Answer-first page structures that are easy for AI to lift, quote and attribute
- Original data, expert insight and first-hand experience: the stuff AI can’t make up (or at least shouldn’t)
- Comparison, “best for”, pricing and FAQ content for the prompts that drive decisions
- Content refresh programmes so your facts, prices and specs stay current
- Writing and editing by humans who know New Zealand markets
Digital PR to build consensus
Get the internet agreeing about you.
AI models triangulate. They look at what many sources say about you and go with the consensus, and your own website is only a small part of that. McKinsey found a brand’s own sites make up just 5–10% of the sources AI search references (McKinsey, 2025). Ahrefs’ study of 75,000 brands found YouTube mentions correlated more strongly with AI visibility than any other factor, while link volume showed only weak correlation (Ahrefs, 2026).
- Earned coverage in NZ media, trade publications and the “best of” roundups AI engines love to cite
- Review, directory and marketplace profile optimisation
- Presence where the conversation happens. Wikipedia, YouTube and Reddit are the most-cited sources in Google’s AI summaries (Pew Research, 2025)
- Shaping the story others tell, so the strengths you want AI to repeat show up in independent sources
AI visibility tracking & analysis
Measure what matters (and ignore what doesn’t).
AI answers are famously fickle. SparkToro’s research found there’s less than a 1 in 100 chance that ChatGPT or Google’s AI will give the same list of brands in any two of 100 responses to the same prompt (SparkToro, 2026). So anyone selling you an “AI ranking” is selling fiction. We track visibility share for you and your competitors across large prompt sets run on an ongoing basis.
- Share of voice and citation share across ChatGPT, Google AI Overviews and Gemini (and others as required)
- Accuracy and sentiment checks: is AI describing you correctly, and kindly?
- Competitor benchmarking by topic and prompt cluster
- GA4 set-up to isolate AI-referred traffic, leads and revenue
- Monthly analysis that comes with actual recommendations, not just charts
AEO vs SEO: what’s the difference?
From Google Search’s perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO.
Google Search Central, 2026Google has a point. Most of what makes AEO work is good SEO done properly. But “most” is doing some heavy lifting there, especially once you look beyond Google to ChatGPT, Perplexity and friends. Tap the diagram to explore:
Where AEO and SEO overlap
- Crawlable, indexable, fast pages
- Helpful, expert content that answers real questions
- A clear, consistent brand entity (with structured data to support it)
- Experience, expertise, authority and trust (E-E-A-T)
- Strong reviews and brand reputation
- Smart information architecture and internal linking
Mostly SEO territory
- Ranking positions and blue-link click-through rate
- Targeting keywords by search volume
- Link building to grow authority
- SERP features: image, shopping and local pack results
- Search Console as the main scoreboard
Where AEO goes further
- Being mentioned and cited inside AI-generated answers
- Conversational prompt and query fan-out research
- Off-site consensus across media, reviews, Reddit and YouTube
- AI crawler access and JavaScript-free rendering
- Influencing what AI models already “know” about you from training data
- Probabilistic tracking across many repeated prompts
How to rank in AI search
“Rank” is a slightly generous word for something this probabilistic, but here’s what consistently moves the needle:
Let the right bots in
Allow AI search crawlers, and make sure key content loads without JavaScript.
Make sure AI knows it’s you
One consistent set of facts, a clear category and an unambiguous name, so every mention connects back to you.
Keep doing great SEO
AI assistants lean on search results to build answers. Rank well there, and you give yourself more chances to be cited.
Answer real questions, early
Put clear, direct answers near the top of the page, then back them up in depth.
Bring proof
Statistics, quotes, sources and first-hand expertise make content more quotable.
Get talked about
Earn mentions in the media, reviews, videos and communities AI trusts.
Track many prompts, not one
Measure visibility share over time, not one lucky screenshot.
Get your AI Visibility Snapshot
Wondering what ChatGPT says about you when you’re not in the room? We’ll run real-world prompts for your category and show you:
- Where you’re mentioned, and where you’re missing
- Which competitors AI recommends instead
- Whether AI recognises your brand, and what it gets wrong
- What we’d fix first
AEO FAQs
Is AEO replacing SEO?
No. Google still sends websites roughly 190 times more traffic than ChatGPT does (Ahrefs, 2026), and AI assistants lean heavily on search indexes to build their answers. AEO builds on a strong SEO foundation rather than replacing it. The best results come from running the two together.
How long does AEO take to work?
Technical fixes can show up within weeks once AI crawlers revisit your site. Content and digital PR usually take a few months to build the consensus AI relies on. Engines that search the live web (like ChatGPT search, Perplexity and AI Mode) respond faster than answers drawn from a model’s training data, which is updated less often.
Can you guarantee my brand will appear in ChatGPT?
No, and be wary of anyone who says they can. AI answers are probabilistic, so the same prompt can return different brands every time. What we can do is increase how often, and how accurately, you appear across a representative set of prompts, and prove it with data.
Why does ChatGPT get facts about my business wrong?
Usually for one of three reasons. The model is relying on what it learned in training, which can be months or years old. The sources it finds disagree with each other, so it picks whichever version it trusts most. Or your name is close enough to another business, place or word that it blends the two together. The fix is entity optimisation: one consistent set of facts across your website and the profiles AI checks, clear signals about who you are (and who you aren’t), and regular monitoring to catch errors before they spread. Our guide to entity optimisation explains how it works.
Should I block AI crawlers?
It depends on your goals, but blanket-blocking usually costs you visibility. Many AI companies run separate bots for model training and for search. OpenAI, for example, uses GPTBot for training and OAI-SearchBot for ChatGPT search (OpenAI). That means you can often opt out of training while staying visible in AI search. We’ll help you set the right rules for your business.
Do I need an llms.txt file?
Probably not yet. Google says you don’t need special AI text files or markup to appear in its generative AI features (Google). An llms.txt file is cheap to add, but we prioritise the things with proven impact first.
Does schema markup help with AI search?
Structured data helps machines understand your entities, products and facts, and it still earns its keep in classic SEO. But Google is clear that it isn’t required for generative AI search and there’s no special “AI schema.” In AEO terms, schema is one input into how clearly AI recognises your brand. Consistent facts across the web and independent mentions do more of the heavy lifting. Treat it as good hygiene, not a magic switch.
How do you measure AI visibility?
We run large sets of real-world prompts repeatedly across the major AI engines and measure visibility share, citation share, sentiment and accuracy, benchmarked against your competitors. We also separate AI-referred traffic in GA4, so you can connect visibility to leads and revenue.
How will AI search affect my website traffic?
Expect fewer clicks on some informational searches. Pew found people clicked a traditional result on 8% of visits when an AI summary appeared, versus 15% without one (Pew Research, 2025). The visitors who do arrive from AI tend to be well-informed and ready to act, though. Adobe found AI-referred retail traffic converted 54% better than non-AI traffic (Adobe, 2026).
Why work with an AEO agency rather than doing it in-house?
AEO spans technical SEO, content, PR and analytics, and it changes month to month. Most in-house teams have one or two of those skills, not all four. We bring the whole mix, over 20 years of search experience with New Zealand brands, and tracking that’s ready from day one. We’re also happy to work alongside your team and upskill them as we go.
Sources & further reading
- 1News: Kiwis send 10 million ChatGPT messages a day (OpenAI data), Sep 2026
- FutureFive: IAB Australia & Pureprofile Commerce & Discovery Report 2026 (NZ)
- IAB NZ: Kiwi Search Habits 2025
- TechCrunch: ChatGPT reaches 900M weekly active users, Feb 2026
- Google: Search at I/O 2026
- Ahrefs: AI Overviews reduce clicks to top pages by 58%, May 2026
- Ahrefs: Q1 2026 AI Search Benchmark (75,000 brands)
- Digital Commerce 360: Adobe AI-referred retail traffic, Jun 2026
- Adobe: AI traffic grows but retail sites lag in AI visibility, Apr 2026
- McKinsey: New front door to the internet, Oct 2025
- Bain & Company: Consumer reliance on AI search results, Feb 2025
- Pew Research Center: Google users click less when an AI summary appears, Jul 2025
- SparkToro: AIs are highly inconsistent when recommending brands, Jan 2026
- Search Engine Land: ChatGPT performs a search in 31% of prompts (Nectiv), Oct 2025
- Aggarwal et al.: GEO: Generative Engine Optimization (KDD 2024)
- Vercel: The rise of the AI crawler, Dec 2024
- Google Search Central: Optimizing for generative AI features
- OpenAI: Overview of OpenAI crawlers