When AI Stops Answering Questions and Starts Doing the Work
Something shifted in how AI works for businesses over the past six months, and the numbers suggest most New Zealand companies haven’t yet noticed. We spent 2023 and 2024 getting comfortable with AI tools that answer questions, summarise documents, and generate text on command. That version of AI was genuinely useful. But the frontier has moved, and the technology now entering commercial use operates in a fundamentally different way.
The term circulating in technology circles is agentic AI. Rather than waiting to be asked something and returning a result, an agentic system takes a sequence of actions to complete a complex task — checking data sources, making decisions, updating records, and reporting back — without a person directing every step. The distinction matters more than it might sound. A chatbot helps you think. An agent does things on your behalf.
Global investment in this technology has accelerated at a pace that is genuinely difficult to comprehend. Worldwide AI investment reached $242 billion in the first quarter of 2026 alone, up from $59.6 billion in the same quarter of 2025. That is a fourfold increase in twelve months. Gartner now forecasts that 40 percent of enterprise applications globally will be integrated with task-specific AI agents by the end of this year, up from less than five percent in 2025.
New Zealand’s position in this picture is uncomfortable. A joint study by KPMG and the University of Melbourne placed New Zealand among the least AI-confident and least AI-trained workforces in the developed world. Only 41 percent of Kiwi workers report using AI at work in any meaningful way. While 87 percent of New Zealand businesses are using AI in some form, only 12 percent have scaled it beyond isolated pockets. The gap between where the global market is heading and where New Zealand currently sits is not narrowing — it is widening.
This is not a piece intended to generate alarm. The purpose of writing about this is to make the practical implications clear, because what is actually changing has direct consequences for how New Zealand businesses should be thinking about their digital infrastructure.
Take a very recent local example. On 22 April, Employment Hero launched HeroForce AI in New Zealand as part of a global rollout. HeroForce is designed to address the compliance burden that has plagued New Zealand small and medium businesses for years. The Holidays Act alone has triggered hundreds of millions of dollars in remediation costs across the economy over the past decade. HeroForce deploys an agentic system that includes an HR agent providing policy-aware guidance, a recruitment agent that handles candidate screening and conducts AI-led interviews, and a payroll agent in development that will monitor real-time compliance. The system draws on live government data and regulatory integrations to stay current.
The point worth noting here is not that Employment Hero built an impressive product, though they have. The point is the type of product. This is AI that plugs into existing business workflows, connects to live regulatory data, and acts on your behalf within a defined authority boundary. It does not wait to be asked. It monitors, interprets, and intervenes. That is what agentic AI looks like in practice, and it is arriving in New Zealand whether businesses are ready or not.
For those of us working in web development and digital strategy, the implications of this shift are significant. The role of a business website has already evolved from brochure to transaction engine to customer service interface. The next evolution is as an integration layer — the public-facing surface of a network of automated agents and systems working behind the scenes. Businesses that are already integrating AI agents into their back-end operations will increasingly look to connect those systems to their customer-facing digital presence. What does your website look like when an AI agent can query your inventory in real time, personalise content based on a visitor’s history, handle a complex product enquiry from start to quote without a human in the loop, and update your CRM the moment a conversation ends? These are not speculative future scenarios. The building blocks exist today, and the businesses beginning to put them together are doing so on WordPress, Shopify, and custom web platforms built with current tooling.
For smaller New Zealand businesses, the practical path forward does not require a wholesale transformation. The advice I consistently give clients at this stage is to start with one process — the task that consumes the most time and creates the most friction — and test an agentic workflow against it. Measure honestly. Tools like Zapier AI and Microsoft Copilot Studio now allow relatively accessible entry points into agentic automation without requiring a development team. For anything more sophisticated, particularly where the agent needs to interact with a live website or customer database, you will need technical help, but the cost of entry has dropped significantly.
What matters most right now is not having the most advanced AI system. What matters is understanding what these systems can do, where they fit in your operations, and what your website needs to become to support them. Businesses that develop that understanding now are building a foundation. Those waiting for the technology to settle down before engaging are watching from an increasingly distant vantage point.
The window for New Zealand to close the gap with global peers is not closed. But the data on investment rates and adoption trajectories suggests it will not stay open indefinitely. Agentic AI is not a future trend to monitor. It is a present-tense shift that is already reshaping what competitive digital infrastructure looks like — and the question for New Zealand businesses is not whether to engage with it, but how quickly.
Roger McSaveney is General Manager of Forge Creative, an Auckland web design and development agency.
Have you started exploring AI agents in your business or on your website? Share your experience in the comments below.