In response to persistent crop insurance frauds, the Uttar Pradesh government has recently enlisted the expertise of four private agencies, employing Artificial Intelligence (AI), Machine Learning (ML), and Remote Sensing (RS) technologies. These companies are tasked with compiling and updating crucial agricultural data, covering land use, crop coverage, crop health, and yield estimates in the state.
This move comes in the wake of a major fraud discovery, where approximately 8,000 farmers allegedly claimed crop insurance benefits without actually planting crops or through deceptive inflation of cultivation area. Beyond financial losses, such fraudulent activities hinder the settlement of legitimate crop insurance claims for farmers who suffered from natural calamities.
Advertisement
Fraud, in its various forms, plagues the entire insurance industry, including health, life, general, and crop insurance. According to Deloitte's Insurance Fraud Survey 2023, approximately 60 percent of respondents perceive a significant increase in fraud, which is largely the fallout of growing digitalisation and weakened controls to prevent or detect fraud.
For instance, policyholders may exaggerate damage to inflate claim values, while identity theft allows fraudsters to purchase policies and submit fraudulent claims. The health insurance sector, in particular, grapples with fraudulent billing for services never rendered. As digitisation advances, new fraud trends like data mining, data theft, and mis-selling of insurance products have emerged as significant concerns.
Advertisement
Tech Transformation Gets Underway
While technology-driven innovation has brought agility and convenience to the insurance and insurtech sectors, it has also exposed vulnerabilities. AI and ML have emerged as powerful tools for analysing extensive datasets, identifying anomalies, and flagging suspicious claims based on historical data and patterns.
K V Karthik, partner for Forensics, Financial Advisory at Deloitte India, underscores the evolving nature of technology and its impact on controls: "With frauds becoming a board-level agenda and digital boundaries constantly blurring, there is a clear need for insurers to relook at their operating model that integrates a larger agenda, which will work across business, compliance, legal, underwriting, and operations departments."
The Indian insurance industry has embraced AI and ML swiftly for various purposes, spanning underwriting, claims settlement, and customer experiences. The India Fintech Report 2022 highlights the rapid growth of AI, ML, IoT, automated claims processing, e-commerce insurance marketplaces, web aggregations, and software/white label/APIs. It predicts that embedded insurance using technology will drive substantial consumer growth in the near future.
Moreover, there is a growing clientele for niche products, like micro, on-demand, and bespoke insurance plans, which are customised to individual needs. This has prompted companies like Artivatic.AI and Mantra Labs to harness NLP, AI and ML technologies.
Advertisement
For instance, Artivatic.AI offers ML and AI-based algorithms for insurers and consumers, offering risk-based personalized automated solutions for underwriting, claims, risk, and fraud intelligence.
AI, on the other hand, can analyse photos and videos to detect signs of fraud and monitor policyholder behaviour to identify suspicious changes. ML-backed chatbots and virtual assistants can engage with policyholders during the claims process, identifying potential fraud and enhancing the customer experience.
A case in point is Plum's PolicyGPT chatbot, based on the OpenAI GPT-3 architecture and NLP model to generate human-like text. It offers users information about their health insurance policy. Saurabh Arora, chief technology officer and co-founder of Plum, believes that technologies like PolicyGPT can bridge the gaps that insurance companies currently face in terms of customer service and response time.
Advertisement
AI for Informed Decision-Making
Leveraging AI and ML empowers insurtech companies to make data-driven decisions when underwriting policies. This allows them to price premiums and offer tailored coverage options accurately. The result is a win-win situation, where customers benefit from fairer pricing while companies can more effectively mitigate risks.
Srinivasan Parthasarathy, MD and CEO of Digit Life Insurance, points out that the diversity of languages spoken in India presents a unique challenge when training AI models. Still, it's an issue that can be overcome with appropriate linguistic data. He notes, "The models will eventually need to be trained in NLP and automatic speech recognition using huge sets of various linguistic data to make AI truly accessible to all Indians."
Advertisement
Many insurtech startups in India face the formidable task of implementing AI models from the outset. However, they encounter two significant challenges—limited data and concerns about data quality. These obstacles require innovative approaches to overcome.
Sarvendu Singh, Head of Engineering at Onsurity, emphasises the importance of effective deployment and prudent management of AI models, even under financial constraints. Integrating advanced AI solutions with legacy infrastructure necessitates thorough planning and adaptability.
With a hint of optimism, Singh notes, "It's precisely these challenges that drive our determination to innovate, adapt, and revolutionise the insurtech landscape."
Building Trust In AI
AI's growing integration into different insurance processes, including underwriting, pricing, claims, and distribution, will reshape the insurance sector in the coming years. However, the trust factor remains a significant challenge.
Advertisement
Insurtech players use AI to predict trends and models based on empirical data, but a lack of historical data or siloed datasets can hinder accurate interpretation. Kartik Ayalh, head of technology delivery at MassMutual India, underscores the importance of ensuring customer confidence through maturing AI with guard rails.
As AI continues to redefine the insurance industry, stakeholders are exploring ways to deploy AI and ML applications that align with their business objectives, making the sector more adaptive and resilient.