Artificial intelligence (AI)-driven insights and machine learning (ML) increasingly play a pivotal role in strengthening fraud detection in life insurance claims, protecting the integrity of the risk pool, says JSE-listed financial services group Discovery Life.
This, as the insurance industry deploys the technology to flag anomalies, tailor health interventions and streamline the processing of claims.
While fraud represents less than 0.5% of claims, the insurance group says advanced analytics and AI tools enable it to flag unusual activity quickly, guide investigations and prevent potential losses.
Discovery Life says it paid a total of R11.5 billion to clients in 2025, with R6.9 billion consisting of individual life insurance claims, R2.4 billion in shared-value payments and R2.2 billion in group risk claims.
In an interview, Sylvia Steyn, head of life claims and service at Discovery Life, tells ITWeb that by combining AI with human judgement, the company can quickly settle low-risk claims, while ensuring high-risk or unusual claims receive the scrutiny needed.
This dual approach protects clients, safeguards the risk pool and enhances the efficiency of claims operations, she says.
“Discovery Life employs in-house systems that combine AI, ML and data analytics to identify suspicious claims.
“Analytics are used to proactively identify potential claims to be paid to customers, as well as to enhance fraud prevention. Claims investigations rely less on external high-tech tools and more on in-house systems, data analytics and human judgement. Sophisticated fraud prevention is to the advantage of customers, since it protects the risk pool but also allows low risk claims to be identified and processed quicker.”
KNOW MORE:
Want to see how AI is being applied in real businesses? Join industry experts and decision-makers at ITWeb AI Summit 2026 on 22 April in Johannesburg, where leaders will unpack practical AI strategies, real-world case studies and the technologies shaping tomorrow’s intelligent enterprises.
According to the Insurance Fraud Bureau, despite the insurance industry preventing R1.4 billion in fraud in 2024, insurance fraud costs the South African insurance industry billions annually, with estimates suggesting short-term insurance claim fraud exceeded R3.5 billion, while life insurers lost over R131.6 million in 2024.
Steyn explains that Discovery Life’s combined AI and ML technologies pull data from multiple sources, including the Department of Home Affairs, fraud databases and industry trends, creating alerts that help assessors decide whether further investigation is needed.
The technology underpins claims assessment and identification of potential claims, helping the insurer manage over 9 000 claims per month.
In most cases that are flagged as suspicious, the insurer’s teams may involve on-site visits or verification with treating doctors, or refer claims to a medical panel for review, she notes.
This approach allows low-risk claims to be processed efficiently, while ensuring potentially fraudulent activity is caught early, reducing unnecessary delays for legitimate clients.
Personalised health insights
Beyond fraud prevention, AI is increasingly used by Discovery to improve clients’ health-related outcomes through personalised health recommendations.
According to Steyn, Discovery Life’s Personalised PayBack Booster, introduced in 2025, rewards clients for completing health actions recommended through Discovery Health’s Personal Health Pathways (PHP).
“PHP uses artificial intelligence to analyse each client’s health profile and engagement patterns to recommend personalised preventative actions and screenings. Within a year of launching this feature, over 18 000 additional screening actions were completed by 14 775 clients, including nearly 9 500 cancer screenings and over 5 400 health checks.”
Early detection through these AI-driven interventions enabled timely treatment and triggered 47 severe illness benefit claims, including 32 early-stage cancers and 12 heart and artery conditions in 2025, she states.
In terms of providing support and assistance during the claims processes, technologies such as predictive analytics, graph-based social network analysis and generative AI tools help assessors process claims more efficiently, extracting information from unstructured documents, and screen risk accurately.
“AI enables assessors to process claims more efficiently and with greater insight,” Steyn says. “Predictive analytics and other advanced techniques, such as graph-based analytics used in social network analysis, support the identification of potentially suspicious claims for further investigation.
“Generative AI can extract and structure information from unstructured data sources, including scanned documents, making it quickly accessible to both models and assessors.
“All AI-driven processes adhere to strict data protection standards and relevant legislation. Clients can choose policies delinked from Vitality data, but around 85% opt for linked policies, reflecting confidence in integrated benefits and AI-enabled services.”
Share
