New Zealand businesses urged to redesign work before adding AI

SAP New Zealand managing director Adrian Griffin says the country’s next phase of technology investment is shifting from cloud migration towards data and AI, but businesses should resist simply adding AI tools to existing processes.

Speaking at the New Zealand SAP User Group in Auckland, Griffin argued that organisations should first rethink workflows, roles and where human judgement adds value. AI and automation can then be applied to the parts of the process where they genuinely improve the way work gets done.

Key facts

  • SAP sees data and AI as an important next phase of technology investment for large New Zealand organisations.
  • Griffin says AI adoption should be treated as business transformation rather than a technology project.
  • Organisations should redesign workflows before introducing AI use cases.
  • Human oversight and clear accountability should remain part of AI-enabled processes.
  • SAP expects agentic AI and automation to change traditional technology support and service models.
  • New Zealand’s technology skills shortage could constrain organisations trying to carry out this work.

Our take

This is a useful distinction for New Zealand organisations starting to move beyond individual AI tools. Buying Copilot, ChatGPT or another platform does not itself create an AI operating model.

The better starting point is often the work itself. Identify the decisions people make, the information they need, the repetitive steps that could be automated and the points where human judgement must remain. Governance can then be designed around the actual process rather than added after the technology is deployed.

Sources

About the author

Campbell McKenzie is a Director at Incident Response Solutions, a New Zealand firm experienced in cyber incident response, digital forensics, investigations and technology risk. Through KiwiGen.AI, Campbell helps professional services firms adopt generative AI safely, with practical governance and controls.