EXL, a data analytics and digital operations and solutions company, has introduced the EXL Insurance LLM, a specialised Large Language Model (LLM) tailored for the insurance industry.
The model, developed in collaboration with NVIDIA AI Enterprise (NVIDIA), is designed to improve key insurance functions, including claims reconciliation, data extraction, anomaly detection, and summarisation, marking a shift towards more customised AI tools for the sector.
EXL said the EXL Insurance LLM addresses industry-specific challenges, which standard LLMs have struggled to meet.
The company said that generic models often lack the fine-tuning required for complex insurance tasks, such as claim adjudication, leading to inefficiencies, extended settlement periods, and increased indemnity costs.
By incorporating proprietary data and a deep understanding of the insurance domain, EXL’s model aims to mitigate these issues, providing more accurate and consistent outcomes.
A Gartner report forecasts that by 2027, over half of AI models used by enterprises will be industry-specific, compared to just 1% in 2023.
EXL claims its LLM has shown a 30% accuracy improvement over top pre-trained models like GPT-4 and Claude.
Developed by EXL AI Labs, the model leverages NVIDIA’s full-stack AI platform, incorporating NVIDIA NeMo for customisation, low-rank adaptation (LoRA), and supervised fine-tuning during the training process.
The LLM uses the NVIDIA Triton Inference Server for enhanced GPU efficiency, and includes NVIDIA NeMo Guardrails for input-output management, aiming to enhance overall accuracy and performance.
Vilas Madan, EXL’s senior vice president and head of insurance for Australia and New Zealand, highlighted the model’s relevance to the Australian market.
“The EXL Insurance LLM delivers 30% greater accuracy and 30% lower costs than generic LLMs, by not only accelerating claims adjudication by AI-enabling tasks like summarisation, question and answering, but also reducing indemnity spends, while ensuring full regulatory compliance to the new AI laws applicable to Australia,” he said.
The EXL Insurance LLM offers a variety of specialised functions, including the aggregation of structured and unstructured data, contextual classification, and support for real-time customer interactions and provider negotiations. It is expected to expand its role within the insurance industry, with future applications in underwriting, premium audits, and other areas.
EXL has launched the new model as AI adoption grows in the insurance industry.
A recent survey from UserTesting highlighted consumer perspectives on AI in insurance across several markets, including Australia.
Conducted by Talker Research, the survey explored preferences for AI versus human advisors in providing insurance guidance. It found that while the majority of respondents prefer human interaction, AI is gaining ground as a potential tool to simplify complex insurance processes.
In Australia, 25% of respondents believe AI could help them better understand insurance-related information, and 45% see value in AI for comparing different insurance plans.
