APRA AI remarks put risk data under the spotlight

Speaking at the Insurance Council of Australia’s annual conference in Sydney on 6 November 2026, APRA Member Suzanne Smith outlined both the promise and the caution around AI in insurance. The regulator’s message was not that insurers should avoid AI, but that governance, accountability and transparency need to keep pace as models become more influential in underwriting, claims handling, fraud detection and customer service.

The point that matters most for small home service providers is segmentation. If insurers can assess risk with greater precision, they may be better able to distinguish between different types of businesses that previously sat in broad categories. A cleaner working in low-risk residential settings, a gardener using powered equipment, a handyman doing occasional work at height and a technician giving advice in clients’ homes may all look quite different once more data is brought into the underwriting process.

That creates opportunities and pressure. Better information may help well-managed businesses explain their controls, training, contract terms and claims history more clearly. At the same time, businesses with vague service descriptions, outdated sums insured or undocumented subcontractor arrangements may find questions becoming more detailed at renewal. Public contact, hazardous materials, vehicle use, off-site tools and advice-based work can all affect how insurers view public liability insurance and related covers.

For operators who use brokers, the development reinforces the value of being able to describe the business accurately rather than relying on last year’s policy schedule. A useful review may include current services offered, revenue split by activity, staff and subcontractor arrangements, tools and equipment values, any AI or software used in customer communications, and evidence of safety procedures. None of this guarantees easier access to cover or lower premiums, but it can reduce avoidable ambiguity.

The limitation is that AI-driven underwriting is still evolving. APRA’s remarks also point to wider questions about fairness, explainability and whether increasingly precise pricing could leave some risks harder or more expensive to insure. For home service businesses, the practical response is not to predict the technology, but to keep business information accurate, records organised and insurance discussions grounded in what the business actually does.