FWD expands AI capabilities to transform insurance operations

FWD Group, a pan-Asian life and health insurer serving approximately 34 million customers across 10 markets, has placed artificial intelligence at the core of its transformation strategy, as part of its effort in changing the way people feel about insurance. Established in 2013 and listed on the Hong Kong Stock Exchange (1828), the company offers life and health insurance, employee benefits, Shariah and family takaful products, as well as general insurance across Hong Kong SAR, Macau SAR, Thailand, Cambodia, Japan, the Philippines, Indonesia, Singapore, Vietnam and Malaysia.

Embedding AI across underwriting, claims and fraud detection

In 2021, FWD introduced its Smart Insurance Framework, a roadmap to embed AI across operations and simplify customer journeys. Since then, it has built in-house AI models to support underwriting, claims management and fraud detection. By 2024, these systems were delivering tangible results, from sharper risk profiling to faster claims processing and improved fraud detection.

AI-assisted Smart Operations, piloted in Hong Kong SAR, was developed to address delays and rising fraud cases identified in 2022. More than half of claims then took over four days to process, while fraud was driving up costs and eroding trust. The initiative combines predictive analytics and machine learning to streamline claims and uncover suspicious activities.

The AI underwriting model generates real-time risk scores to improve straight-through processing of applications. The AI claims model analyses profiles and histories to identify low-risk claims for automatic approval while flagging high-risk cases for manual review. Fraud detection has been enhanced through anomaly detection and graph network analytics, which reveal suspicious links between customers, providers and agents.

Enhancements in 2024 included replacing rule-based claims checks with the XGBoost methodology, which improved accuracy by flagging two-thirds of high-risk cases previously missed. Transparency was strengthened through SHapley Additive exPlanations (SHAP), which provides claims officers with interpretable insights into risk factors behind each decision.

Tangible results and expansion plans

Results have been significant. By mid-2025, the AI claims model achieved 97.2% accuracy in identifying low-risk personal accident and hospitalisation claims, lifting straight-through processing rates and cutting turnaround times. Fraud detection improved from under 1% to 52%, with seven new fraud, waste and abuse patterns identified. Customer experience metrics also rose: FWD’s net promoter score increased to 63 as of June 2025, while 92% of customers rated their purchase experience positively.

The initiative has freed staff to focus on complex, high-value tasks, supported by AI-generated insights. FWD said this balance of automation and human judgment has raised both productivity and job satisfaction.

Expansion is underway. AI-assisted Smart Operations will be introduced in in other markets, supported by FWD Brain, the group’s centralised machine learning platform. The system standardises data pipelines and accelerates model deployment across markets, with a focus on real-time monitoring, feedback loops and governance to ensure compliance and trust.

FWD’s approach has drawn industry recognition. The company was named winner of Best AI Powered Claims Initiative – Asia at The Digital Banker’s Global AI Innovation Awards 2025, for its innovative use of predictive analytics and fraud detection algorithms to boost efficiency, accuracy and customer service.