NZ’s Next Xero Could Come From The AI Layer That Everyone Dismissed

When ChatGPT launched, the consensus was that all the value would accrue to the foundation model companies like OpenAI, Anthropic, and Google. Everything else would be dismissed as “just a wrapper", a simple user interface built on top of GPT, offering little more than what the underlying model already provides. It was viewed as fragile, easy to copy, and with no real way to defend its value or build a lasting business.

Then came the ChatGPT Store, which only amplified the fear. If OpenAI controlled the platform where most users discover AI tools, independent apps could be crushed overnight. It was the Amazon risk: Amazon monitors which third-party products sell well, launches its own cheaper version, and buries the original offering. Many assumed OpenAI would do the same with wrappers, replicating successful products and cutting out smaller players. Not to mention, every few months, OpenAI rolls out features that quietly kill entire categories of AI apps such as plugins, research tools, and note-takers.

These tools started out basic and some have evolved to deliver tremendous value to users. The real opportunity isn’t in building foundational models, it’s in the application layer. That’s where AI becomes seamless, usable, and truly valuable.

Compare this to the SaaS era. Calling a startup an “AI wrapper” is like calling early SaaS companies “SQL wrappers". It would've missed the point. The real value was never in the database itself but in the layer that transformed the data into polished, intuitive products underpinned with logical workflows.

Think about the giants of the SaaS era: Salesforce did not invent databases. Shopify did not invent payment gateways. And Xero didn’t invent accounting. They built intuitive layers, removed friction, and owned the customer experience.

AI wrappers, when done well, follow the same pattern, turning raw model outputs into products people actually want to use, just as SaaS transformed basic databases into billion-dollar businesses.

If wrappers are proving their value globally, the question is how can New Zealand seize this momentum?

New Zealand will not build the next ChatGPT. We do not have the capital or global-scale datasets to compete with OpenAI or Google. That does not matter.

Our opportunity is at the application layer. According to Microsoft, AI applications and data centres together will generate $3.4B for Aotearoa by 2035, with applications alone creating $2.1B dollars in new revenue. The report is clear - the biggest economic opportunity lies in AI-powered applications built for our core strengths in agritech, fintech, healthtech, tourism, and logistics.

And if the application layer is where we have opportunities, we should lean into it and start simple with AI Wrappers that are purpose-built and differentiated. Wrappers are the fastest way for Kiwi entrepreneurs to test ideas, capture niche markets, and spin up solutions that solve real problems. From there, some will naturally evolve into deeper platforms with data moats and automation layers.

Let's be clear. Many wrappers will fail. Some will be leapfrogged by larger players, swallowed by the rapid innovation coming from foundational models, or collapse because they lacked a roadmap.

Building enduring enterprise value is difficult. Not every wrapper will become a billion-dollar company. But that isn’t the point. For New Zealand, these startups or innovation pockets within existing businesses are critical because they:

When productivity is under fire and growth is scarce, why wouldn’t New Zealand experiment in this space of innovation?

We’re already seeing signs of its success globally. A wave of companies are spinning up with meteoric, headline-grabbing rises. Some will be acquired, some will fade quickly, and others will evolve into strong platforms as smart founders turn simple AI wrappers into enduring businesses with well-defined roadmaps.

Mao, the infamous Chinese leader, once coined the phrase “let a hundred flowers bloom” to encourage ideas to compete, with the best rising through market forces to improve overall output and quality. Despite its controversial origins, the philosophy applies to innovation: encourage broad experimentation and let the strongest ideas win. New Zealand should adopt the same mindset. A surge of AI wrapper experiments (even if most fail) will spark breakthroughs, fuel industries, and may create the next Xero-level success.

The loudest criticism of wrappers is always the same. “There’s no defensibility". This is true to an extent. Many wrappers are at risk of being wiped out by the relentless wave of innovation from LLMs and the big players.

But I disagree with the idea that wrappers are doomed. AI wrappers in the AI age can build moats through (and not limited to).

We are in a platform shift. Every gold rush is noisy and full of failures. But the winners are always those who experiment quickly, move fast, and carve out a niche before the dust settles.

NZ’s next Xero will not come from building models but from turning AI into products people love. Let a hundred experiments bloom, and those winners might define our next decade of innovation.

So, the question isn’t whether AI Wrappers will last forever. It’s whether we’re prepared to use this moment as our launchpad.