KAMAU BONIFACE KATHITHU | The rising threat of insurance fraud in SA: leveraging technology for defence

The insurance industry in South Africa plays a critical role in safeguarding the financial well-being of individuals and businesses. However, it faces the persistent challenge of financial fraud, which poses a significant risk to insurers. As technology advances and fraudulent methods evolve, insurance companies in South Africa face escalating risks made worse by the current economic hardships. People are more inclined to engage in fraudulent activities in such challenging economic conditions.

The fiscal impact of insurance fraud remains challenging to detect and assess. According to the insurance crime bureau, an estimated 20% of the R35bn short-term insurance payouts in 2019 could have been fraudulent, translating to a loss of R7bn in a year. Life insurance companies have also observed a rise in fraudulent say, with the Association of Savings and Investment South Africa reporting an increase of 12% between 2019 and 2020.

Insurance fraud affects all stakeholders within the sector, including insurers, policyholders, beneficiaries, intermediaries, the government and society at large. Effective fraud detection and prevention measures can lead to lower premium rates, improved profitability for insurers, increased competitiveness, enhanced consumer trust, and improved stability and reputation for the insurance sector. Protecting genuine customers and their say is crucial in reducing insurers’ churn rate, significantly improving profitability.

To combat fraud, insurance companies have implemented fraud detection and prevention strategies. Some insurers have established in-house forensic teams to identify and investigate fraudulent activities. However, technology can play a vital role in augmenting insurers’ capabilities in the fight against fraud. Technological tools such as Artificial Intelligence (AI), the Internet of Things (IoT) and Blockchain have effectively detected and prevented fraud.

Machine learning (ML), a branch of AI, empowers insurers to analyse vast data sets and identify patterns that indicate fraudulent behaviour in real-time. ML algorithms continuously learn from historical data, enabling them to identify similarities between past fraudulent actions and classify say as fraudulent or non-fraudulent. These algorithms recognise anomalies and promptly generate alerts when suspicious behaviour is detected in claims data. They can identify unusual say, forged documentation, incorrect personal data and other fraudulent indicators that would be difficult and inefficient to detect manually.

ML algorithms are flexible and can be trained on diverse data types, including speech, text and video. As ML algorithms evolve and become more sophisticated, they can even suggest new risk indicators, facilitating the identification of emerging fraud types.

Digitising the claims process can act as a deterrent to insurance fraud. The advent of IoT has created a globally interconnected network of smart devices. Using IoT technology in the claims processed serves as a powerful deterrent against false claims, as it reduces the window of opportunity for fraudsters to manipulate data in their favour. By leveraging IoT technology, insurers can access and analyse data stored in the memory of intelligent on vehicles or home sensors. For a motor claim, this data could include detailed information about the vehicle’s operation, such as its speed, location, acceleration and braking patterns. By comparing this data with the information provided by the customer during the claims process, insurers can gain valuable insights into the circumstances surrounding the incident.

Chatbots can be used to speed up claims processing. Customers can notify a claim using a chatbot with no human intervention required. The chatbot will direct customers to capture photos and videos of the damage they report. This proactive approach serves multiple purposes, primarily aimed at minimising the opportunity for fraudsters to manipulate data while providing insurers with an advantage in detecting fraudulent activities.

Chatbots encourage customers to document the damage through visual media to ensure the information is more reliable and accurate. Photos and videos serve as tangible evidence of the reported incident, allowing insurers to assess the extent of the damage and make informed decisions regarding the claim. Additionally, this documentation creates a time-stamped record, reducing the likelihood of fraudulent claims being filed for pre-existing damage or unrelated incidents.

The swift collection of visual evidence through chatbots also limits the opportunity for potential fraudsters to alter or exaggerate the extent of the damage. Fraudulent individuals often aim to exploit insurance policies by inflating say or misrepresenting the circumstances surrounding the incident. By encouraging customers to document the damage immediately, chatbots help deter fraudsters who rely on the ability to manipulate data over time.

Insurers can leverage advanced technologies, such as computer vision and machine learning algorithms, to analyse the submitted photos and videos for any inconsistencies or signs of fraud. These technologies can detect image alterations, recognise patterns and compare the visual evidence against existing databases to identify potential red flags. By integrating these fraud detection capabilities with chatbot systems, insurers gain an advantage in efficiently identifying suspicious claims and taking appropriate actions to mitigate fraud.

Blockchain technology offers significant potential in combating fraud within the insurance industry. Blockchain provides an immutable and transparent ledger where transactions and data cannot be altered retroactively. By recording policyholder information, claims data and transaction history on a blockchain, insurers can create a trustworthy and auditable record of events. This transparency reduces the likelihood of fraudulent activities, as any attempts to manipulate or forge data can be identified by comparing it with blockchain records.

Blockchain enables smart contracts, which are self-executing agreements with predefined rules and conditions. These contracts automatically execute actions when specific needs are met. In the insurance context, intelligent contracts can automate claims processing, verifying claim details against predefined conditions and triggering payouts accordingly. This automation reduces the risk of human error and manipulation, making it harder for fraudsters to exploit the claims process.

Blockchain facilitates secure data sharing among multiple parties, such as insurers, policyholders, healthcare providers, and law enforcement agencies. This allows the seamless exchange of information, while ensuring data integrity and privacy. By sharing relevant data on the blockchain, insurers can enhance fraud detection capabilities by cross-referencing information across different entities. This can help fight fraud committed through identity theft. Blockchain-based records can also prevent double-dipping fraud, where someone files a claim with multiple underwriters.

While technology offers significant advancements in detecting and preventing insurance fraud, it should be viewed as a complementary tool rather than a standalone solution. Emphasising the collaboration between technology and human expertise is crucial for insurers to stay ahead in the fight against insurance fraud.

Kamau Boniface Kathithu is an actuary at The Shard with a career spanning over seven years in the insurance sector. His areas of expertise include reserving, capital modelling, product development and reinsurance.