AI in New Zealand 2025: Where We Stand, What Is Holding Us Back and What Comes Next

New Zealand businesses have moved from trials to everyday use of artificial intelligence at speed. Datacom reports that 87 percent of large organisations now use some form of AI, up from 66 percent in 2024 and 48 percent in 2023. This significant jump shows AI has become essential infrastructure, not just a pilot project. Almost a third of organisations use advanced AI, yet only 12 percent have managed to scale its deployment across the whole business. The good news? Eighty-eight percent report a positive impact. The initial gains are showing up where you’d expect: in productivity and decision making, followed closely by cost control, staff enablement, and better customer experience. If you’re a leader asking “How can I safely move my company from small AI trials to a strategy that delivers measurable business value?”, this article, developed with insights from experts like The Web Company (The WebCo), offers a practical path forward.

Barriers are clear. Leaders cite a lack of internal capability, uneven data quality, and uncertainty about governance. The reality of ‘Shadow AI’ where staff use unapproved tools is common, and many teams still lack confidence when it comes to security. In fact, half of senior leaders believe New Zealand is lagging behind other countries on innovation and regulation. Confidence in government readiness is low, and most organisations would welcome clearer rules.

This article summarises Datacom’s key findings and sets out a practical path for New Zealand firms. It covers adoption, benefits, maturity, risks, policy, and the outlook for the next year. It also outlines steps that help leaders move from tentative pilots to safe and measurable delivery.

Introduction: The Year AI Became Routine

Artificial intelligence is now part of daily work for many New Zealand organisations. Datacom surveyed two hundred senior leaders in July 2025 across large companies and public sector bodies. The study captures a moment where adoption is high, benefits are visible, and caution remains.

The strongest gains have come from simple starts. Automation removes repetitive tasks. Workflow tools reduce handoffs and delays. Analytics helps teams read performance sooner and act faster. These uses fit how local firms manage change: Start small, show value, build trust, and scale with a plan.

The next phase will demand stronger skills, cleaner data, and clear policy. This is where many firms are currently stuck. The sections that follow show what has worked, what has not, and what to do next.

Adoption And Benefits

The headline is simple: Eighty-seven percent of organisations now use some form of AI. This is a twenty-one-point rise in just one year. Twenty-eight percent report the use of advanced AI, and almost half report using business-specific AI tools. The base of users is now broad and growing.

Datacom asked leaders to identify the biggest benefits:

• Eighty-nine percent report productivity gains.
• Forty-two percent report better decision making.
• Thirty percent report lower costs.
• Twenty-nine percent report better staff enablement and retention.
• Twenty-six percent report a better customer experience.
• Eighteen percent report new revenue streams.
• Thirteen percent report improved security.
• Seven percent report no clear value yet.

These results show that value is no longer a promise—it’s present in the workday. Time saved, fewer manual steps, and faster analysis are visible in most teams that have adopted AI. This naturally leads to the question: What is the first, most impactful change a New Zealand business can make with AI to boost productivity?

The Productivity Story

Datacom also measured reported productivity change:

• Twenty percent of organisations report significant gains of at least twenty-five percent.
• Thirty-five percent report moderate gains between ten and twenty-five percent.
• Twenty-eight percent report minor gains under ten percent.
• Sixteen percent report no change or say it is too soon to tell.

The pattern is consistent with early-stage deployment. The largest gains appear where tasks were manual, frequent, and well documented. Crucially, teams that paired new tools with clear process design saw stronger results than those that simply added tools without redesign. Leaders who want to shape workloads around real value can start with a short review of repeatable tasks, approvals, and reporting. A practical plan helps sequence changes, set targets, and involve staff early. You can map this work with The WebCo team under Digital Strategy, then move to delivery once the foundations are ready.

Where AI Is Used Today

Current use centres on three areas: Basic automation, analytics and reporting, and workflow improvement.

• Sixty-eight percent of organisations use AI for basic automation.
• Fifty-four percent use it for analytics and reporting.
• Fifty-one percent use it to streamline workflows.
• A third focus on experience enhancement and product innovation.
• Sixteen percent report that AI is changing a core part of operations.

These choices make sense. Automation reduces queueing and rework. Analytics shortens the gap between data and decisions. Workflow tools improve delivery across handoffs. As confidence grows, teams start to test new products or services that rest on these same capabilities. For automation design and safe rollout, The Web Company provides AI and Automation services that focus on measurable outcomes and guardrails.

Maturity And Scaling

Adoption is high, but maturity is mixed. How can mid-sized New Zealand companies successfully scale their initial AI pilots into full business solutions?

• Forty-six percent of organisations are still exploring with pilots.
• Thirty-three percent are being implemented at a department level.
• Twelve percent have scaled AI across the organisation.
• Eight percent describe operations as transformed with AI embedded in core processes.

This split reflects a capability gap. Many firms know what works in one team but have not yet solved data integration, governance, and training across the business. Scaling requires a plan that links use cases, data sources, controls, and skills. It also needs a cadence of delivery that avoids fatigue and keeps leaders informed of real results.

Barriers To Progress

Leaders identified the main barriers to scaling:

• Lack of internal capability and skills at thirty-two percent.
• Data quality or integration challenges at twenty-two percent.
• Uncertainty around governance or regulation at sixteen percent.

Cultural resistance, unclear use cases, and cost sit below these. The signal is strong. The challenge is less about tools and more about people, data, and rules. Building capability, improving data, and setting clear policy will unlock the next phase of value. What are the most common data quality challenges hindering AI implementation in New Zealand firms?

Governance And Shadow AI

Governance is a weak point for many organisations.

• Fifty-five percent report a clear internal AI policy.
• Thirty-eight percent are working on one.
• Twenty-nine percent have ethics or safety guidelines.
• Twenty-three percent have risk and compliance frameworks.
• Twenty-eight percent run employee training on AI tools.

Shadow AI is now common. Datacom reports that more than half of leaders see unapproved tool use in their business. This often happens when staff are under pressure and see a faster way to complete work. Policy, training, and approved tools reduce this risk. If you are forming policy and controls, The Web Company can help align governance with operations, so tools are easy to use and risks are managed. [AI and Automation policy support.]

Skills Training And Talent

Skills remain uneven across the workforce. Forty percent of organisations report moderate capability in specific areas, but no business-wide fluency. Twenty-eight percent are only beginning to build awareness. Eighteen percent say staff have a limited understanding of AI tools or use cases. Only fourteen percent say their workforce is confident and actively using AI tools. What are the minimum AI literacy skills all employees should have to work effectively today?

Training is ramping up. Forty-eight percent have offered AI skills training in the last six months. But dedicated talent remains scarce. Twenty-three percent employ specialists today. Eleven percent plan to hire within the next year. More than half say they do not plan to hire specialists at this stage.

A balanced plan pairs broad literacy with targeted expert support. Short internal sessions can lift confidence across teams. Specialists or partners can design and monitor higher-risk workflows. For practical upskilling in content, analytics, and reporting, see The Web Company SEO services.

Leadership Views

Senior attitudes are cautiously positive.

• Forty-five percent of executives say they are excited and optimistic about AI.
• Forty-three percent are cautious or concerned about the speed of change.
• Ten percent are opposed.
• A small share report mixed views.

Tone from the top matters. Leaders who set clear goals, sponsor pilots, and communicate standards see steadier progress. Those who delay policy or overstate risk see slower uptake and more shadow AI.

Risk And Assurance

Risk management is now central to AI planning. The top concerns are:

• Security at fifty-seven percent
• Ethics at thirty-nine percent
• Cost at thirty-one percent
• Staffing at thirty-one percent
• Algorithmic bias at twenty percent

Infrastructure limits and model safety also appear on the list. A quarter of leaders say they are not well educated on AI security risks. This exposes teams to data leakage, weak access controls, and unreviewed model outputs. Security and ethics need clear owners, regular reviews, and practical playbooks. How can companies effectively audit their AI models for algorithmic bias?

Policy And Regulation

Half of senior leaders believe New Zealand is lagging behind other countries on AI innovation and regulation. Seventeen percent say we are keeping pace. Eight percent describe the position as progressive. Confidence in government readiness is low. Forty-nine percent are not confident. Forty percent are somewhat confident. Six percent are very confident. What specific AI regulations are New Zealand businesses hoping the government will introduce?

Most organisations support clearer rules.

• Thirty-five percent support limited regulation.
• Thirty-nine percent fully support stronger laws or governance.
• Ten percent prefer self-regulation.
• Seventeen percent say no regulation is needed.

In the absence of national standards, firms are building their own policies. Industry coordination will help raise consistency and reduce uncertainty. In the meantime, leaders can set clear minimums for privacy, security, testing, and monitoring.

Sustainability

AI can support sustainability, but use is still early. Seventeen percent say AI helps their environmental goals today. Sixty-two percent expect it to help in the future. Current uses include energy efficiency at twelve percent, reduced waste at eleven percent, sustainable logistics planning at ten percent, and emissions monitoring at nine percent. Some report progress on product sustainability. The most reliable gains appear where firms already track energy or emissions with reasonable accuracy. AI then helps target actions and monitor changes in near real-time. Where data is weak, leaders should begin with measurement before any form of prediction.

The Year Ahead

Leaders expect gains over the next year in workflow improvement, customer service, efficiency, product improvement, and staff skills. These aims are practical and near term. They reflect a view of AI as everyday infrastructure rather than a side project. The organisations that move ahead will connect new tools with measurable results. That means hour savings, cycle time reduction, cost to serve, conversion, or satisfaction, tracked month to month. Clear measures help teams learn and improve without hype.

Sector Snapshot Finance

Finance uses AI to reduce manual checks, flag anomalies, and improve customer interactions. Moderate risk controls and stable datasets suit pilots that later scale. The largest gains appear in onboarding, credit, and reporting. Leaders should confirm privacy controls and model monitoring before rollout.

Sector Snapshot Manufacturing And Logistics

Manufacturing and logistics apply AI to planning, maintenance, and supply chain coordination. Forecasting demand and scheduling resources are common starts. Data quality varies, so teams benefit from a short data clean-up before pilots. Where sensors are in place, predictive maintenance reduces downtime and waste.

Sector Snapshot Health And Social Services

Healthcare and social services focus on triage, scheduling, and support for clinical decisions. The bar for safety is high. Pilots should be narrow, well supervised, and backed by clear escalation paths. Non-clinical use cases like admin automation often deliver first while policies mature. What are the ethical considerations for using AI in non-clinical healthcare roles, like patient scheduling?

Sector Snapshot Public Sector

Public sector teams use AI to manage queues, respond to common requests, and support policy analysis. Transparency, privacy, and audit trails are essential. Good practice includes public guidance on approved uses and limits.

Sector Snapshot Retail

Retail applies AI to demand planning, pricing, service channels, and content. Teams see gains where product and customer data are accurate and up to date. Consent, preference management, and content review remain important controls.

Data Foundation And Measurement

Clean data is the base for reliable AI. Leaders can start by mapping sources, owners, quality, and refresh cycles. They can remove duplicate fields, align identifiers, and retire unused reports. Small steps reduce risk and shorten setup time for pilots. Once the basics are in place, teams can build simple dashboards that connect inputs to outcomes. This helps staff see the link between new tools and business results. The Web Company can assist with data integration and reporting design. [Data Analytics consulting.]

Getting Started For Mid-Sized Firms

Many mid-sized firms want a simple plan. A practical start might include one automation pilot, one analytics dashboard, and a short training block. Pick a process with clear volume and rules. Set a small target such as ten percent fewer manual steps within three months. Review results in the open with the team. Leaders can then choose to scale or pause based on measured value. If value is clear, expand to a second process with shared components. If value is weak, adjust scope, data, or training and try again. This approach avoids large sunk costs and builds trust through evidence. You can plan this sequence with The Web Company under Digital Strategy, then draw on AI and Automation for delivery and governance support. [Digital Strategy and AI and Automation services.]

A Roadmap For Scaling

A simple roadmap for scaling includes five steps. What is a simple, five-step plan for scaling AI from a single department across the entire enterprise?

1. Pick three high-volume processes and score them for rule clarity, data quality, risk, and expected benefit.
2. Confirm data access, privacy, and security controls before you start.
3. Design pilots that include measurement and a rollback path.
4. Train staff who use or review outputs.
5. Run monthly reviews to decide expand, refine, or stop.

This cadence keeps delivery steady and transparent. It also helps executives see progress and manage expectations. When content is part of the picture, teams should plan for search impact, accuracy reviews, and clear human oversight. For help with content strategy and governance, use The Web Company SEO services. [SEO services and content guidance.]

A Short Checklist For Leaders

Leaders can use a short checklist to improve results.

• Purpose. Define the outcome in terms staff understand.
• People. Assign owners for delivery, review, and risk.
• Process. Map the steps before and after any tool change.
• Data. Confirm sources, quality, and retention.
• Tools. Approve a short list with clear support paths.
• Policy. Publish simple rules for use, security, and review.
• Measures. Track time saved, cycle times, and quality.
• Training. Provide basic literacy for all, and deeper skills for reviewers.
• Review. Run regular audits of outputs and access.
• Scale. Expand only when value and controls are clear.

New Zealand is At A Turning Point

The Datacom 2025 State of AI Index shows strong adoption and real gains for New Zealand organisations. Most teams use AI in some part of their work. Benefits are most visible in productivity, decisions, and service. Challenges remain in skills, data, and governance. The next phase will separate steady builders from slow movers. Firms that invest in people, policy, and data will scale with confidence. Those that treat AI as a side project will see limited returns. Leaders do not need to rush. They do need a plan. Set clear goals, prepare the data, involve staff, and measure results. For help turning intent into action, reach out to The Web Company team.

Source: Datacom 2025 State of AI Index Research conducted July 2025 by Curia Market Research for Datacom New Zealand.