While general insurers in Malaysia recognise the potential of predictive AI and generative AI for claims decision-making and fraud detection, General Insurance Association of Malaysia’s Mr Chua Kim Soon says that challenges to adoption within claims functions include fragmented data, legacy processes and capability gaps. Despite this, significant progress in digitalising motor claims process has been realised.
Malaysia’s motor insurance sector is facing similar challenges to its Southeast Asian peers, especially in rising repair costs, claims frequency and the changing risk profile brought about by electric vehicles (EVs) and advanced vehicle technologies.
For instance, speaking with Asia Insurance Review, General Insurance Association of Malaysia (PIAM) CEO Chua Kim Soon highlighted that in 2025, Malaysia’s private-car claims frequency remained above 7%, while average claim severity increased 20% to RM8,831 ($2,183).
Said Mr Chua, “Overall, the motor segment recorded a combined ratio of 103%, indicating that claims and expenses continued to exceed premiums.”
Using AI in claims
Touching on claims, Mr Chua cited WTW’s 2025 Malaysian Claims Analytics Survey, done in collaboration with PIAM saying, “While insurers recognise the potential of predictive AI and generative AI for claims decision-making, fraud detection and EV claims assessment, adoption within claims functions remains relatively low due to fragmented data, legacy processes and capability gaps.
“This suggests Malaysia is at an important transition point, rather than necessarily being ahead or behind the region overall.”
Digitalising claims
“Malaysian general insurers have made significant progress in digitalising motor claims process, although some manual touchpoints remain due to legacy systems, established processes and involvement of multiple stakeholders across the claims ecosystem, including workshops, adjustors, tow trucks providers and other service providers,” said Mr Chua.
Additionally, insurers must consider data, he added. “Claims information is often generated and held across different platforms, making it important to enable faster verification, assessment and decision-making.
“But at the same time, certain claims still require human expertise, particularly for more complex damage assessment, fraud detection and cases where supporting documentation or verification is required.”
In the case of EV claims, Mr Chua also pointed out that these considerations are “becoming more specialised”.
“Battery expertise, diagnostics, high-voltage systems, ADAS calibration, specialised parts and repair expertise can require additional assessment and access to appropriately equipped and certified repair facilities,” he said.
“The availability of technical expertise, parts and repair infrastructure can therefore have an impact on claims turnaround times.”
Going forward, he believes the focus will be on “further reducing unnecessary manual touchpoints through greater digital integration, automation and data connectivity, while ensuring that human expertise remains available where it adds value”.
“This will be particularly important as EV adoption increases and motor claims become increasingly technology- and data-driven,” said Mr Chua.
Factors limiting adoption
There are several factors that limit the adoption of advanced analytics in motor and EV claims.
The first is data availability and quality. “Claims data is often generated across multiple systems and industry stakeholders, with differences in data formats, completeness and accessibility,” Mr Chua said.
This, he added, can make it challenging to develop reliable analytics models and risk profiles.
“For EVs, the relatively limited volume of claims data also makes it more challenging to build predictive models, given the relatively new and evolving nature of the technology,” he said.
“EV risk assessment also depends on the maturity of the wider repair ecosystem, including battery diagnostics, calibration capabilities and access to specialised repair expertise. These capabilities are still developing as EV adoption grows.”
Other considerations he listed included:
- Availability of specialised analytics and technical talent.
- Investment and implementation costs.
- Data governance, PDPA and regulatory or compliance requirements.
Despite these challenges, he said, “These factors are part of the industry’s ongoing digitalisation journey rather than fundamental barriers to adoption.
“As the ecosystem matures and more data becomes available, insurers will be better positioned to develop more robust analytics capabilities and apply them effectively across motor and EV claims.”
Building models for claims
“The Malaysian general insurance industry is increasingly recognising that industry-level data sharing is important because individual insurers may not have sufficient EV claims volumes to build reliable predictive models on their own,” said Mr Chua.
This is where PIAM steps in, he continued, as the association “provides an important platform for industry collaboration, including through the collection and dissemination of industry data and engagement with regulators and other stakeholders”.
“For motor and EV claims specifically, the industry is looking at strengthening the collection of EV-specific information and developing a more consistent data pool covering areas such as claim frequency and severity, repair costs, battery-related damage, parts availability and repair turnaround times,” he said.
According to Mr Chua, PIAM has also “highlighted the need for collaboration with original equipment manufacturers, technology partners and certified workshops to build EV claims capabilities”.
“This is particularly important as EV claims data remains limited,” he said. For instance, WTW’s report also noted data sources as a barrier, and called for unified claims data architectures.
Complex claims
The human element in claims must also be balanced against automation, especially in more complex claims.
“The objective should be to improve efficiency without compromising customer confidence, transparency and service quality, particularly in complex EV claims involving battery assessments or specialised repairs,” he said.
“In the Malaysian context, the ideal approach is to automate processes, and not empathy. AI and digital tools can efficiently handle routine activities such as accident reporting, document submission, claim status tracking and preliminary damage assessment, making the claims process faster and more convenient for customers.”
He also stressed that human involvement “remains important for complex or sensitive motor and EV claims, particularly those involving serious injuries, disputed liability, total-loss decisions, battery damage assessment or exceptional customer circumstances”.
As he put it, these cases “require professional judgement, fairness and empathy, where human interaction remains valuable”.
“Ultimately, the appropriate approach for Malaysia is a human-in-the-loop model, where AI and digital tools enhance efficiency, consistency and decision support, while claims professionals retain oversight and handle complex decisions and customer interactions where human judgement and reassurance are most needed,” he said.
Looking into fraud
“A key concern is that fraud is becoming more sophisticated and technologically enabled. In Malaysia, common frauds such as staged accidents, inflated repair or battery claims, counterfeit parts and manipulation of digital evidence, are evolving as technology advances,” said Mr Chua.
“This highlights the need for insurers to continuously strengthen their fraud detection capabilities.”
He also stressed that for EVs, “the risk profile is also increasing as EV adoption rises”.
“The higher value and complexity of components such as batteries, sensors, software and advanced driver assistance systems, could create opportunities for fraud,” he said.
“As such, insurers will need stronger EV-specific data and expertise to distinguish legitimate high-cost claims from potentially fraudulent claims.”
Keeping up
According to Mr Chua, to keep pace with fraud, insurers are “strengthening industry collaboration, data sharing and claims intelligence to identify emerging fraud patterns and strengthen detection capabilities”.
“Data analytics and industry-wide claims intelligence can help insurers identify anomalies and patterns that may not be apparent when claims are assessed individually,” he said.
“The development of a comprehensive motor claims database, supported by Bank Negara Malaysia, could further strengthen the industry’s ability to identify emerging fraud patterns.”
He also pointed out that overall, the focus “should be on building the necessary data, technology and governance capabilities now, so that the industry is better prepared to prevent EV-related fraud as the market grows”. A