When AI changes how boards assess value

AI is already changing how organisations operate before many boards have made a major AI investment decision.

Staff are using it to speed up everyday tasks. Suppliers are adding it to software and services organisations already rely on. Management teams can point to pilots, productivity tools and early gains. In some organisations, AI is already shaping analysis, customer contact or board material before directors have had a full conversation about its role in the business or service model. 

The first phase has been relatively easy. People have experimented, found useful shortcuts and tested AI in parts of the business where the risk felt manageable.

The next phase asks more of boards. Activity is easy to find. Progress is harder to prove. 

The question has shifted from AI awareness to AI value. The issue is whether AI is changing the parts of the organisation that determine performance, and whether the board has enough evidence to support the investment and organisational change involved.

A useful board test is whether the business or service model would look the same if the organisation were starting now, with AI available in ordinary tools, systems and services. 

The 2025 IoD Director Sentiment Survey shows AI and digital oversight moving into mainstream board work. Six in 10 boards are working with management on how AI and automation can lift productivity. The same survey points to uneven assurance, with cyber vigilance plateauing over the past three years.

AI value depends on confidence in the systems and people behind it. It also depends on cyber and data settings keeping pace with how AI is being used. A board may hear a strong productivity story while still needing a clearer view of how a tool is being used, what data it relies on, how outputs are checked and who remains accountable.

AI Forum New Zealand’s 2025 report, AI in Action: New Zealand’s Third AI Productivity Report, found that 91% of surveyed businesses reported efficiency improvements from AI. Half reported positive financial impacts and 77% reported lower operating expenses. The figures support the case for AI use, and raise the next board question: are the gains confined to doing existing tasks more quickly, or is AI changing a constraint that limits performance?

Faster work can be valuable, especially in organisations where time and capacity are scarce. The stronger productivity case comes when AI changes a bottleneck in the business or service model: the way customers are served, decisions are prepared, advice is produced, risk is detected or scarce expertise is used. 

The answer will differ by organisation. A large organisation may be considering whether AI changes customer service, core decision processes or the economics of a major function. It may also need stronger assurance over uses that affect people or sensitive information. A smaller business, charity, Māori entity or public-purpose organisation may be focused on a more immediate constraint, such as scarce expertise or slow decisions. In each case, directors need to understand the problem AI is meant to change. 

Treasury’s 2025 work on innovation, capital and productivity reinforces a familiar point. Technology lifts performance when it is matched by capability, investment and changes to how the organisation operates. AI will follow the same path. A list of pilots shows effort. A clear account of what is changing shows whether the effort is becoming value.

The investment conversation is becoming more demanding. KPMG’s 2026 Global AI Pulse found that AI is moving rapidly across organisations, while enterprise performance is still catching up with investment. Some returns will take time because the early activity is often below the surface: preparing data, connecting systems, building staff capability and putting assurance around the tools.

Directors do not need instant proof for every investment. They need to know what evidence would justify continuing, what would point to narrowing the work, and when management would redirect effort. 

AI is also entering the board’s own information flow. Diligent’s 2026 research report, What Directors Think, found that 66% of boards reported using AI in the boardroom, while only 3% had fully integrated AI into risk oversight and strategic decision-making.

A board paper may still have a named author, an executive summary and a recommendation. However, the work behind it may have changed. AI may have summarised research, compared options, drafted the risk section or turned rough analysis into confident prose. Used well, it can improve the paper. Used carelessly, it can make weak analysis look more settled than it is.

Verifying AI-assisted material now belongs in routine board practice. Where AI has contributed to material coming to the board, directors need to understand what it was used for, which sources it relied on, what a person checked and where judgement remains uncertain.

The assurance task becomes more important when AI moves from supporting a person to triggering the next step in a process. An agentic system may send an alert, open a job or move information between systems. The National Cyber Security Centre’s 2026 guidance on agentic AI places these systems inside the cybersecurity environment, with a focus on secure deployment, access control and monitoring.

Directors need to understand what authority has been given to the system, what it can access, how management will detect unexpected behaviour, and when intervention is required.

AI also changes organisational dependencies. Many New Zealand organisations will rely on AI tools owned, hosted or developed offshore. Directors need a clear view of where data is processed, how supplier terms treat organisational material, and how the organisation would respond if an AI-enabled process failed, became unavailable, changed suddenly or became materially more expensive.

The pace of adoption will depend on trust and capability. KPMG’s New Zealand insights from its global AI study found that only 24% of New Zealanders surveyed had undertaken AI-related training or education. Only 36% believed they had the skills to use AI tools appropriately.

Policy helps, but capability carries the work. People need to know what to put into a tool, when to check the answer and when to ask for help.

New Zealand’s AI Strategy and MBIE’s Responsible Artificial Intelligence guidance give boards a useful foundation, including privacy, cybersecurity and human oversight. The boardroom task is to turn those principles into choices about investment, service, people, assurance and accountability.

Effective AI governance depends on connecting AI use with the organisation’s business or service model, and with the assurance directors need to trust the change. Directors are asking what is changing, what evidence supports the investment, which systems the organisation now depends on and where accountability sits.

AI creates value when it helps the organisation do important things better, or do things that were previously difficult or uneconomic. The board's role is to test whether AI is making existing activity more efficient or changing the business or service model in ways that justify further investment and organisational change.

Exclusively for IoD members

Click below to download the AI agents of change reader, the second of the Top 5 issues for directors in 2026. You will be prompted to log in before accessing a downloadable PDF.  

Download reader