EY India has developed a ‘Customised Fine-Tuned LLM’ tailored specifically for Banking, Financial Services, and Insurance (BFSI) sector, promising to revolutionize customer service and operational efficiency. As the industry moves towards vertical LLMs, EY India’s tailored LLM offers advanced AI-driven capabilities. It enhances answer accuracy, intent recognition, contextual comprehension, and BFSI-specific vocabulary depth, empowering clients with more precise and dependable AI outcomes.
Designed with a deep understanding of localized customer service with terminologies and regulatory requirements, the model offers multilingual support, including English and Hindi. This advanced AI solution leverages the fine-tuned LLAMA 3.1-8B instruct model to address the distinct challenges of customer care in the BFSI industry. The model delivers measurable business outcomes, including up to 50% cost savings.
This BFSI sector focused fine-tuned LLM has been trained specifically on Indian BFSI domain data and understands the contextual vocabulary and questions better than some of the most popular foundational LLMs available in the market, proved using scientific evaluation frameworks. This is possible due to extensive understanding of the domain inserted into the LLM via Parameter Efficient Finetuning with Low Rank Adaptation techniques.
Mahesh Makhija, Partner and Technology Consulting Leader, EY India, said, “”AI is rapidly transforming industries and driving economic growth. EY’s customised AI models are tailored to meet the specific needs of the BFSI sector, empowering banks to optimize operations, enhance risk management, and make data-driven decisions. By integrating AI into their daily business, banks can unlock new revenue streams, improve financial resilience, and contribute significantly to the broader economic landscape, reinforcing India’s position as a leader in the AI-driven economy.”
EY’s customised fine-tuned LLM can be hosted either on-premises or in a cloud environment with limited GPU requirements, ensuring significantly lower costs. BFSI organizations can leverage this model for customer service applications via text channels such as WhatsApp, SMS, and website chatbots, as well as voice-calling channels within their AI call center frameworks.
Rohit Pandharkar, Technology Consulting Partner, EY India, said, “The BFSI sector in India is setting a remarkable precedent in AI adoption. By leveraging customised AI models like EY’s fine-tuned LLM, the sector is not only enhancing customer care with deeper contextual understanding and regulatory compliance but also driving innovation in operational efficiency. This pioneering approach underscores the sector’s commitment to leading the way in responsible AI adoption.”
Unlike other models, this fine-tuned solution mitigates issues such as hallucinations and regulatory risks by being securely hosted within enterprise networks, eliminating concerns around data privacy and cross-border data exposure.
