EXL, a leading data analytics and digital operations solutions company, has unveiled the EXL Insurance LLM, a new large language model (LLM) designed to transform critical tasks in the insurance sector, such as claims and underwriting.
This new product was developed in response to the ongoing challenges insurance companies face when trying to leverage general-purpose LLMs.
These models often lack the ability to process complex, industry-specific tasks, leading to inefficiencies such as higher indemnity costs, claims leakage, and increased compliance risks.
The EXL Insurance LLM addresses these issues by focusing exclusively on insurance-related operations.
The product is designed to handle tasks such as claims reconciliation, data extraction, anomaly detection, and question-answering.
Built by EXL AI Labs on the full-stack NVIDIA AI platform, the LLM delivers 30% greater accuracy on insurance tasks compared to pre-trained models like GPT4, Claude, and Gemini. It also ensures regulatory compliance while significantly reducing costs.
Further enhancing its capabilities, EXL used the NVIDIA NeMo™ platform for training and fine-tuning the model, employing advanced techniques like low-rank adaptation (LoRA) and supervised fine-tuning (SFT).
This solution leverages NVIDIA Triton Inference Server™ to maximise GPU performance and includes retrieval-augmented generation (RAG) to manage long documents and generate insightful responses.
In addition, the EXL Insurance LLM integrates NVIDIA NeMo Guardrails for better control of input and output, which improves the overall user experience. This robust framework makes the product ideal for processing complex medical records, handwritten notes, and claims data.
Anand “Andy” Logani, EXL’s executive vice president and chief digital officer, said, “With 25 years of expertise in processing medical records data for bodily injury, workers’ compensation, and general liability claims, EXL has developed curated data sets with domain-specific tagging, labeling, and question and answer pair creation for claims adjudication to fine-tune our models. The EXL Insurance LLM offers 30% greater accuracy and 30% lower costs than generic LLMs while ensuring full regulatory compliance.”
The EXL Insurance LLM supports a wide array of tasks, including data ingestion, classification, and triaging for claims adjudication, payment integrity, and customer service. By improving negotiation accuracy, detecting anomalies, and offering real-time insights, the LLM empowers insurance providers to enhance their decision-making processes and customer interactions.
Developed by the EXL AI Labs, the EXL Insurance LLM continues to evolve and expand its use cases across the entire insurance value chain, including underwriting, premium audits, and finance, further demonstrating its versatility in addressing industry-specific needs.
Copyright © 2024 InsurTech Analyst
