In this CIO Edge session, CareSuper CTO Simon Reiter shares how APRA and ASIC guardrails shape AI pilots, training, and trust at scale within the organisation.
The fund runs two parallel streams: one focused on governance, ethics and standards, and another exploring practical use cases surfaced through an internal AI think tank.
Staff identify more than 140 opportunities, many relating to efficiency, member engagement and investment analysis.
He stresses that value comes from solving real problems, not from chasing technology.
Caresuper pilots internal knowledge bots, AI‑supported member‑service workflows and enhanced investment research.
Adoption varies across teams, but usage is monitored through an AI dashboard that tracks tools, models and activity.
Co‑pilot is the most widely used platform, while specialist teams also use models like Claude and ChatGPT.
To build trust, the organisation invests heavily in change management.
Executives and managers complete deep‑dive AI training, and internal sessions help staff understand use cases and reduce fear about job impact.
Simon says successful scaling depends on strong data quality, integration capability and clear identity management.
Agents must have distinct identities so their actions can be tracked, audited and reviewed like human performance.
Moving from pilots to production also requires mature governance, robust processes and ongoing evaluation.
He emphasises that AI should free people from low‑value tasks so they can focus on complex work that improves member value, service quality and organisational velocity.
Key takeaways:
- CareSuper approaches AI with strong governance first, ensuring ethics, compliance and regulatory alignment before deploying any agentic tools.
- AI adoption grows through people‑focused change management, with executives and teams trained early so staff feel confident, informed and supported.
- Scalable AI depends on foundations, including clean data, robust integration, and clear identity management so every agent’s actions can be tracked and reviewed like a human contributor.