Weak Systems Threaten Insurer Profits

Southeast Asia’s general insurers risk a reduction of up to four percentage points in profitability if they implement artificial intelligence (AI) on outdated data systems, as limited infrastructure weakens the potential benefits of automation.

Raunak Mehta, co-founder and chief executive of insurance technology firm Igloo, told Insurance Asia that properly directed AI investment over the next three to five years could improve profitability by 300 to 400 basis points. However, he stressed that such gains would not be evenly distributed across the sector.

“If the right investments are made in AI over a period of three to five years, you should definitely see this going up by 300 to 400 basis points,” he said.

He added that insurers failing to modernise core systems may not experience comparable improvements. According to Mehta, many general insurers in Southeast Asia operate with very high combined ratios, leaving them with “wafer-thin net incomes” in many cases.

AI expenditure across the insurance sector is rising. A March report by Boston Consulting Group, Inc. indicates that investment in AI could increase from about 0.6% of revenue in 2025 to 1.9% this year. Despite this growth, only around one-third of property and casualty insurers fully recognise AI’s role across operational workflows, even though it has the potential to reduce costs and support premium growth.

Mehta noted that AI can significantly reduce claims-handling costs, improve early fraud detection, and minimise overpayments. It may also lower loss ratios by enhancing risk selection and reducing adverse selection.

“It will definitely go a long way in bringing down the loss ratio by cutting down fraud and, in a way, doing a better job of reducing adverse selection,” he said.

A separate industry outlook by KPMG International Ltd. shows continued confidence in the sector despite macroeconomic and climate-related risks. Around 82% of insurance chief executives reported confidence in company growth, up from 74% the previous year, while 78% expressed confidence in the broader industry.

However, Mehta cautioned that such optimism must be considered alongside operational constraints. “You need to separate the hype layer from what I call the infrastructure layer,” he said, noting that many insurers still rely on legacy systems that are decades old and require substantial upgrading before AI can be deployed effectively.

Igloo itself processes approximately 90 million policies each month and serves more than 50 million customers across the region, according to Mehta.

KPMG further reported that AI remains a top priority across the sector, with 73% of chief executives identifying it as a key investment area. In addition, 67% plan to allocate between 10% and 20% of budgets to analytics, automation, and generative AI.

Cybersecurity continues to rank as a major concern, with 83% of executives citing cybercrime as the most significant threat. Mehta, however, said that faster systems do not inherently increase risk, noting that many cyber incidents are driven by social engineering rather than system performance.

He emphasised that insurers should prioritise data quality and internal system integrity to improve decision-making capabilities. “While we are not building any foundation models ourselves, what we are focused on is the proprietary data that we have, how we are utilising it to extract enterprise-grade knowledge and decision-making,” he said.

Key Industry Indicators