As adoption grows and AI becomes operationalised, New Zealand businesses are no longer asking, "Should we use AI?" the more mature question now at the forefront is, "How do we prove AI’s value?"
Most can already point to AI use cases working across the business, be it a customer service agent here or an automated expense categorisation there. Each one proves value in its own patch. But proving value in one area doesn't automatically add up to value across the organisation. In the adoption phase, it's easy to end up with a handful of smart, siloed solutions or treat AI as a series of one-off technology projects rather than an enabler of broader business transformation.
So, when customers ask me, "Does my organisation need an AI strategy?", my answer is simply, yes. Not because another document is needed, but because without one, there's little sharing of perspective, opportunities or enterprise-level requirements, or more succinctly, little connectivity to scale, operate and govern AI effectively.
Yes, you need an AI Strategy and the data backs it up
Deloitte's 2026 Global Human Capital Trends report, which for the first time included a dedicated New Zealand view, found that AI adoption is accelerating faster than the organisational work design needed to make the most of it. Just 2% of New Zealand organisations say they are leading in intentionally designing how humans and AI work together, compared with 7% globally.
KPMG and the University of Melbourne's 2025 Trust, Attitudes and Use of AI study found a similar pattern inside New Zealand workplaces: only 44% of employees say their organisation has an AI strategy at all, and the same proportion cited having policies in place to govern how it's used responsibly.
The cost of getting this wrong is real. MIT's NANDA initiative found that 95% of enterprise generative AI pilots fail to deliver measurable profit-and-loss impact. The issue isn't necessarily the technology itself, but that pilots can get bolted onto existing processes rather than connected to a broader plan. KPMG's 2025 CEO Outlook found 43% of New Zealand CEOs expect it to take three years or more to see a return on AI investment, compared with just 14% globally, a sign many are still working out how to turn scattered wins into compounding ones.
Where organisations get stuck
In my role as Technical Director AI and Data at Inde, I see this pattern play out with clients: adoption across New Zealand organisations now sits somewhere around 60–70%, but only around 20–30% have taken the next step of strategising AI at an enterprise level. What many build instead is a prioritised list of use cases and mistake it for a strategy. Useful for sequencing work, but often silent on who owns AI adoption once it's live, or how the organisation manages the risk that comes with it.
That gap gets more consequential as AI moves from answering questions to making decisions. A client recently questioned the need for an AI Strategy. They'd already invested in a data strategy, so why duplicate the effort? The answer: a data strategy tells you your data is clean and available. It doesn't tell you which decisions an AI agent is allowed to make with it, who's accountable when it gets one wrong, or how that risk changes once AI is acting autonomously rather than just retrieving information. One feeds the other; neither replaces it.
The economics are shifting underneath all of this too, in a way that can catch budgets out. AI tools that were free to trial two years ago are now licensed and token-metered, so scaling something that worked in pilot can cost far more than the pilot suggested. And there's a quieter risk alongside the cost: open-source options can be attractive because they're free, but that openness can expose organisational data. Even mainstream paid tools carry risk if sensitive information is pasted into them without a clear policy on what's appropriate to share.
Where an AI strategy proves its worth
An AI strategy isn't a document that sits alongside your technology roadmap for the sake of it. Done well, it does a few specific jobs that individual department projects can't do on their own:
-
It sets a shared foundation. Data quality, governance, security and access controls are the same problems whether it's finance, marketing or customer service asking the question. Solving them once, centrally, is faster and safer than every team solving them separately, or not solving them at all.
-
It creates a common way to prioritise. Not every use case deserves investment, and the ones that get the most attention aren't always the ones that create the most value. A strategy gives leadership a consistent way to weigh initiatives against actual business outcomes.
-
It assigns ownership for scaling, not just building. Plenty of pilots have an owner. Far fewer have someone accountable for turning a working pilot into an enterprise capability. Someone needs to be accountable for AI adoption across the organisation, the tools being used, the guardrails around them and the path to scaling what works.
-
It connects what's already working. The finance team's automation or the customer service agent don't need to be the same tool, but they should be able to sit on the same data, governance and security foundation, so success in one team can be replicated, not reinvented, in the next.
-
It treats AI as a capability, not a project. The technology, the use cases and the organisation will all keep changing. A strategy needs a built-in way to revisit direction, rather than being written once and left on a shelf.
Why this matters
New Zealand organisations already know AI can deliver value. The question is whether those individual wins can become something bigger. Without a strategy, they risk remaining isolated successes, owned by individual teams and difficult to scale.
An AI strategy provides the connection between experimentation and enterprise capability: what to prioritise, how to govern it, who owns it and how to scale what works.
So in answer to the initial question without a doubt your organisation needs an AI strategy. Not another document, but a way to turn AI from a collection of successful projects into an organisational capability.
