Insurers across Asia Pacific are accelerating AI adoption, chasing an opportunity analysts estimate could add US$70 billion in value to the industry. But progress is uneven, held back by fragmented data sources and a patchwork of regulations that vary from country to country.
Where AI is making inroads
Claims processing and underwriting are the two areas where carriers have seen the clearest gains. Several large Japanese and Australian insurers have deployed machine learning models to automate routine claims triage, cutting processing times by more than 40% in some pilot programs.
In life insurance, underwriting engines trained on structured application data are now handling up to 30% of standard policies without human review in Singapore and Hong Kong, according to industry reports. The savings are modest today but are expected to scale as models improve.
The data bottleneck
Data remains the biggest obstacle. Many insurers still operate on legacy core systems where customer, policy, and claims data sit in separate silos. Cleaning and linking that data is a prerequisite for any AI project, and it is proving slower than executives anticipated.
A second problem is data quality. Models trained on biased or incomplete historical data produce skewed outputs, particularly in pricing and risk assessment. Regulators in Australia and India have flagged concerns about algorithmic fairness, and several carriers have paused automated underwriting rollouts pending regulatory guidance.
Regulatory divergence
No unified AI regulation exists across Asia Pacific. Singapore has published a model governance framework that encourages experimentation. China requires AI systems in insurance to be approved by the central bank and the financial regulator. Japan has taken a lighter-touch approach, relying on existing insurance law.
This fragmentation forces multinational insurers to build different compliance workflows for each market, increasing cost and slowing cross-border deployment. Smaller regional players face a steeper burden: they lack the legal teams and compliance budgets to navigate multiple regimes simultaneously.
Why this matters for insurance professionals
For underwriters and claims handlers, the shift means routine decisions will increasingly be automated, but the need for judgement in complex cases will rise. Professionals who understand both the business logic and the limitations of AI models - especially around data quality and bias - will be in demand. Carriers are quietly investing in internal training programs, but most are still early. The opportunity is to build that expertise now, before the regulatory and technical clarity arrives and the competition for talent intensifies.
For more on how these changes affect day-to-day roles, see AI for Insurance.
Your membership also unlocks:
