The number that should stop every AI pitch meeting
PwC’s 29th CEO Survey, covering more than 4,500 chief executives globally, landed a finding in February that deserves more airtime than it has received: 78% of New Zealand CEOs report AI has had little to no impact on their organisations. New Zealand sits worse than the global average on this measure. Half of those same leaders say their greatest concern is not transforming fast enough.
That combination, low impact and high anxiety, is not a technology problem. It is a sales environment perfectly designed to separate businesses from their capital.
The global execution gap is worse than the headlines suggest
KPMG’s Global Tech Report 2026, published in January, puts numbers around the disconnect. While 74% of organisations globally report their AI use cases are delivering some business value, only 24% achieve ROI across multiple use cases. The ambition is even more disconnected: 68% of organisations want to reach the highest level of AI maturity by end of 2026, but only 24% say they are there today.
High performers are getting average ROI of 4.5x. The industry average sits at 2x. Most New Zealand businesses are closer to the average, if they are measuring at all.
Guy Holland, Global Leader of KPMG’s CIO Centre of Excellence, framed the shift: “The future belongs to leaders who turn intelligence into advantage. Our research shows organizations are pushing past the early phase of AI roulette, placing scattered bets on multiple technologies, and are now increasingly focused on delivering value.”
The talent shortage compounds everything. 53% of organisations globally still lack the people needed to deliver their digital transformation plans, and 92% anticipate that managing AI agents will become a critical skill within five years.
NZ’s SME problem is structural, not motivational
The MBIE AI strategy published in July 2025 established the baseline: 67% of larger NZ businesses utilise some form of AI, up from 48% in 2023. But 68% of NZ SMEs had no plans to evaluate or invest in AI, compared to only 38% of Australian SMEs. At publication, New Zealand was the only OECD country without a published AI strategy.
The AI Forum’s March 2025 survey of 200 NZ businesses found 82% reported some AI use, but the detail matters. In 2025, 72% were using pre-existing AI solutions built into tools they already owned, versus only 13% using custom-built solutions. Only 56% reported positive financial impact. More than a quarter of firms deploying AI had done so without structured upskilling.
The firms that get it right are seeing real money
This is not a story about AI being worthless. A report commissioned by 2degrees and prepared by Deloitte Access Economics, published in April 2026, provides the first NZ-specific estimate of AI’s relationship with firm-level productivity. The headline: the average SME that adopted AI earned around $400,000 more in FY25 than a comparable non-adopter. The average large business earned approximately $59.1 million more.
2degrees Chief Business Officer Andrew Fairgray said: “The research shows that AI is no longer theoretical. It actually gives us some practical levers that we can use to lift productivity, if we adopt it and we adopt it properly.”
The critical caveat: this study measures firms that have adopted AI. It does not measure the broader population that invested and saw little return. Self-selection is doing heavy lifting in those numbers.
Treasury’s sobering baseline
A Treasury economic analysis from March 2024 found no evidence of aggregate productivity growth from AI despite positive firm-level impacts in some cases. A 10% increase in AI investment was associated with only 0.04% firm-level growth. The analysis described the economics literature as increasingly pessimistic about productivity gains if adoption is left to market dynamics.
In August 2025, BusinessDesk flagged poor progress monitoring and poor follow-through on initial training as specific execution failures holding NZ firms back.
What this means before your next vendor meeting
The pressure to spend on AI is at maximum precisely when the evidence base for broad returns is weakest. The businesses seeing genuine gains share specific characteristics: integrated data systems, digitised core processes, clear governance, and workforce capability built before the technology arrived.
For every other business, the practical question is simple. Before the next AI pitch meeting, ask the vendor to show you ROI evidence from businesses your size, in your sector, with your level of data infrastructure. If they cannot, you are not buying productivity. You are funding their case study.
Sources
- Execution gap: 78% of CEOs say AI has had little to no impact (2026-02-25)
- KPMG report warns AI scale-up lags lofty ambitions (2026-01-22)
- AI Moves From Hype To Hard Results For Kiwi Businesses, 2degrees Study Reveals (2026-04)
- AI IN ACTION: Exploring the Impact of Artificial Intelligence on New Zealand’s Productivity – Report 2 (2025-03)
- Behind the AI hype: Hidden risks, rising costs and execution failures (2025-08-14)
- Update on the Impact of Artificial Intelligence – an Economic Analysis (2024-03)