Last week we had the pleasure to attend the Westpac Smarts | Building a Smarter Business: Planning for the AI Era meeting, hosted by Business Canterbury.
The meeting focused heavily on the practical business impact of AI adoption and the reality that AI implementation is now shifting from experimentation into operational transformation.
A major theme throughout the discussion was that AI success is less about the technology itself and more about organisational mindset, governance, enablement, and business structure. Several attendees highlighted that simply giving staff access to AI tools will not automatically create efficiencies. Instead, organisations need AI champions, clear prioritisation frameworks, and structured governance to identify high-value use cases and coordinate adoption across teams.
Another key discussion point was the changing commercial model created by AI. Businesses that traditionally sold time-based services are now questioning how to price AI-driven efficiencies. If tasks that once took hours now take minutes, organisations need to reconsider whether they are selling labour, expertise, outcomes, or intellectual property.
Concerns were also raised around whether businesses unintentionally give productivity gains back to customers for free, reducing profitability while increasing client expectations. The group explored how AI lowers barriers to entry and reduces traditional scarcity advantages.
There was strong awareness around risk. Multiple attendees highlighted that poorly governed “homemade AI solutions” could become one of the largest operational and security risks facing businesses. Connecting untested AI tools directly into core business systems without proper security layers, controls, or governance was seen as a significant concern.
Operationally, the meeting touched on emerging AI rollout considerations such as SaaS licensing versus token-based AI agent consumption, the shift from capital expenditure toward operational expenditure, ROI-focused AI investment decisions, composable architecture and reusable AI components, AI governance councils using prioritisation methods such as dot voting, and the role of fractional AI engineers and specialist enablement support.
The overall tone of the discussion was pragmatic rather than hype-driven. The strongest message was that successful AI adoption requires disciplined change management, structured governance, people enablement, and strategic thinking, not simply access to new tools.