AI in insurance: current trends and what’s next?
The benefits of AI have been widely recognized across the insurance sector, with 7 out of 10 insurance company leaders seeing the technology as a top investment technology. Consequently, the AI in insurance market is estimated to anticipate high growth rate in the next decade, estimated at 28.4% annually between 2024 and 2030 and potentially reaching US3.59 billion by 2030 [1]. This momentum reflects the substantial value AI creates across both the industry and enterprise levels. Sector-wise, AI powers significant improvements in core functions, driving up to 40% reduction in costs of onboarding new customers, a 10-15% increase in premium growth, and a 3-5% rise in claims accuracy [2]. On a business level, insurers with mature AI and digital capabilities outperform their peers financially, gaining a 6.1 times higher total shareholder returns (TSR) than their least advanced counterparts [3].
Underwriting and claims are among the insurance functions most transformed by AI. Nearly half of global insurers now rely on AI to streamline these processes, making them two of the most widely adopted applications in the industry [4]. And worldwide companies are already seeing results from this investment with a leading Asian insurance group serving as a success story. The company leveraged AI to automate part of these processes, from using OCR to automatically scan and extract information, to having AI make decisions based on pre-set criteria and suggest payout options. Consequently, the insurer managed to cut processing time for both procedures to 2 seconds per request, down from 36 hours for underwriting and 2 days for claims.
Beyond workflow automation, AI is increasingly shaping the future of underwriting through advanced risk assessment. Insurers are beginning to leverage AI to analyze real-time and unstructured data, from those captured by IoT devices to customer behaviors and lifestyle. According to the Organization for Economic Co-operation and Development (OECD), several insurers are already applying AI in property insurance underwriting and pricing. One example involves the use of machine vision and deep learning to generate 3D models of property attributes and assign risk scores accordingly. Another analyzes data from several sources including earth observation and combine with change prediction analytics to improve pricing accuracy by 20% [5].
In claims, insurers are advancing toward straight-through processing to deliver a fully digital, AI-driven experience for both customers and assessors. One leading insurer has taken the lead by partnering with FPT to build an end-to-end automated, AI-powered claims solution as illustrated below.
The company’s claims process before:
At the same time, AI is playing a growing role in fraud detection by identifying anomalies and inconsistencies in claims submissions. It also enables automated cross-checking of documents and images against available databases. One Asian insurer has successfully applied this approach, combining AI with additional security measures such as encryption, regulatory compliance, and proactive threat monitoring. As a result, the company increased fraud detection rates by 50% and reduced fraud-related losses by up to 15%.
What to watch out for?
1. Legacy systems:
If AI is a well-oiled machine, then data is the oil that keeps it running. Without high-quality, sufficient data, AI delivers limited value, and at worst, biased outcomes. Despite the wealth of data that insurance companies possess, a significant portion is trapped in silos due to the fact that the insurance industry has one of the oldest legacy systems. Unlocking this data requires modernization, which in turn raises the costs of AI adoption.
But here’s a catch: AI requires abundant data and modernized systems to run on, but it is also AI that speed up the process of legacy modernization. AI tools help developers analyze legacy programming, automatically generate replacements, and speed up transformation with higher accuracy. For instance, FPT’s xMainframe can interpret and interact with legacy mainframe systems and COBOL code with up to 97% accuracy. Subsequently, Gartner predicts that by 2027 generative AI like xMainframe will be able to explain legacy applications and generate replacements, potentially reducing modernization costs by as much as 70% [6].
2. Regulation & Compliance:
As one of the most heavily regulated industries, compliance is critical for insurers. Yet the regulatory landscape is far from static, and rules are tightening at an unprecedented pace as AI models advance. This means insurance systems must update at the same pace and given the complexity of these systems, doing so on a regular basis is no easy task. For example, implementing a new regulation typically requires impact analysis, development, testing across individual systems, and integration before going live. This process can take anywhere from two weeks to three months, creating risks of business disruption or, worse, non-compliance.
FPT’s Smart Business Rule (SBR) solution addresses this challenge by enabling rule updates directly on the SBR platform with no programming required. Once approved, rules are automatically synchronized across integrated systems, reducing implementation time to as little as three days to two weeks.
AI-first insurance requires an AI-first partner
The future of insurance is undeniably AI-first. Realizing that vision, however, requires the right AI-first partner to provide a clear roadmap, accelerate delivery, and ensure scalability. With its homegrown AI platform, FleziPT, FPT drives innovation through an AI-driven development approach, award-winning solutions, and a deep talent pool certified by global partners including NVIDIA, Microsoft, SAP, AWS, and Google Cloud. Over more than 20 years , FPT has partnered with global insurers such as AIA, Prudential, Allianz, and more to advance their digital transformation and AI journeys.
Learn more about FleziPT - FPT’s AI-first platform here.
