AI search has moved fast over the last couple of years. I know you're running your business and looking after your clients, and it's hard to keep up with all the things that keep changing in this space.
So it's worth spending five minutes now to read through and understand what those shifts have been. AI engines are answering the kinds of questions your buyers used to type into Google, and the businesses being named in those answers are quietly building a position that's going to be harder to close once the rest of the market catches on.
At a glance
- ChatGPT alone now handles 2.5 billion prompts a day across 900 million weekly users, and most NZ B2B companies have no idea whether they're in those answers.
- Checking yourself doesn't tell you much, because ChatGPT serves you back what it already knows about you.
- A clean audit, built from ICP-derived questions and uncontaminated API calls, gives a far truer picture than checking yourself ever will.
- Most NZ B2B categories aren't being named in AI answers yet, so the first credible voices to show up own the space.
Why has search shifted so fast, and why haven't most NZ businesses noticed?
Buyer behaviour has changed. People aren't scrolling a list of links anymore. They're asking a question and getting an answer, and they're moving from awareness to consideration to decision inside a single conversation with an AI engine.
The space is moving fast. Here's where it's at right now.
The shift in numbers
- ChatGPT: 900M weekly active users (Feb 2026), 125% YoY growth, 2.5 billion prompts a day.
- Google AI Mode: 1B monthly users. Gemini: 900M monthly users, doubled in 12 months.
- 58% of Google searches now end without a click (Penske Media v Google, 2025).
- Google I/O, May 2026: the biggest changes to search in 25 years. Sundar Pichai: "Search has become less about individual queries and feels more like an ongoing conversation."
This is not a trend. This is just how it is now.
Can I just check this myself in ChatGPT?
Yes, you can, and it's worth doing as a first look. Open ChatGPT in an incognito window, log out of your account, and type a question your buyer would ask. You'll get a cleaner answer than if you're logged in, where ChatGPT is biased toward what it already knows about you.
A self-check tells you something. It just doesn't tell you the whole picture.
The audit I run for clients tests hundreds of questions across ChatGPT, Claude, and Gemini, all derived from your ICPs and mapped across the awareness, consideration, and decision stages. That's the difference between a snapshot and a real picture of how the market sees you.
A self-check vs a clean audit at scale
| A quick self-check | A clean AEO audit at scale | |
|---|---|---|
| What it covers | A handful of questions you can think of in five minutes | Hundreds of questions across the buyer journey, derived from your ICPs |
| Bias | Lower if you're incognito and logged out, higher if you're not | None. API calls with no history, no memory, no browser bias |
| Engine coverage | Whichever one you're using at the time | ChatGPT, Claude, and Gemini, run in parallel |
| What you walk away with | A useful starting point | Where you stand in the market, by buyer stage, with the content gaps named |
What did running this on UpliftRFP actually show?
Before recommending Resonance's AEO audit to any client, I ran it on UpliftRFP, the SaaS business I co-founded. I built a set of questions our ICPs would realistically ask the major LLMs, then ran them through the proprietary tool I built to score the answers across ChatGPT, Claude, and Gemini.
The headline numbers came back:
- Overall visibility: 37%
- Decision stage: 83% (buyers who already knew our name were finding us)
- Consideration stage: 15%
- Awareness stage: 3%
The decision-stage score was reassuring. The 3% at awareness was the real signal.
Here's what flipped it from "ouch" to "this is the opportunity." Not many companies were named at the awareness stage at all. What was there was generic. White space, not a crowded field. Every unanswered question the audit surfaced turned into a content brief: authentic, specific, ours to own. We keep seeing the same pattern across the clients we run this for. Clear competitor maps, clear gaps, clear content priorities. The gaps are real, and right now the answers in most NZ B2B categories are still generic.
What should NZ B2B leaders be doing about this right now?
Three things matter most.
- Blog content. Blogs get cited more than anything else. Question-led, written from real experience, anchored to specific numbers and moments.
- AEO-specific technical implementation. Schema markup, JSON-LD entity signals, AI crawler permissions. The things AI engines actually look for when they decide what to pull in.
- Lived-experience content. Real numbers, real client moments, real opinions. This is the catnip. It's what LLMs cite, and it's the one thing competitors can't copy.
And here's the signal that this is already real, not future. A Resonance client recently told us their website is now being indexed by large language models more often than humans are visiting it. That's happening now, in New Zealand. The businesses that are technically and editorially ready to be cited are the ones being carried into AI answers. The ones that aren't, aren't.
The white space is wide open right now. It will not stay that way.
Frequently asked questions about AEO and AI visibility
Is AEO just SEO with new keywords?
No. SEO ranks pages in a list. AEO gets your business named inside a synthesised AI answer. Different content, different signals, different game.
How is a clean AEO audit different from typing my own questions into ChatGPT?
Scale. A clean audit runs hundreds of ICP-derived questions across ChatGPT, Claude, and Gemini via API calls with no memory, no chat history, no browser bias. You get a comprehensive picture of where you stand by buyer stage, not a one-off snapshot.
Why does this matter now if AI search is still new?
Because most NZ B2B categories are wide open. The companies that get named in AI answers now build a compounding citation advantage, and the ones that wait will be playing catch-up against competitors the AI engines have already settled on.