Generative artificial intelligence (AI) — and its most popular model ChatGPT — have begun making deeper imprints on the banking, financial services and insurance (BFSI) industry.

Generative AI models, as the name suggests, are trained extensively on large datasets to quickly generate new content in text and multi-media formats on user prompts. Given this, their potential use-cases in banking and finance can be numerous.

That’s particularly true of ChatGPT, powered by the advanced GPT-3.5 architecture — a large language model (LLM) invented by OpenAI and integrated into the New Bing (Microsoft’s search engine). It possesses a remarkable capability to understand and respond to natural language inputs.

The use cases of ChatGPT for the BFSI industry cover, but are not restricted to:

• Conversational banking: Developing chatbots that can promptly address customer queries, saving time and enhancing customer experience
• Fraud prevention: Analysing customer behaviour and identifying patterns indicating potential fraud, enabling BFSI companies to take proactive measures
• Risk analysis and synthetic data: Examining financial data and generating reports that highlight potential risks and opportunities, assisting BFSI companies in informed decision-making
• Personalised financial advice and AI research assistants: Providing tailored recommendations for financial products and services based on customer profile, thereby improving customer engagement and loyalty
  

Leading financial institutions have already put LLMs to work. This trend is expected to extend much further:

• JPMorgan Chase claims it employs LLMs for analysing legal documents, automating report generation, and enhancing customer service through chatbots
• Morgan Stanley is collaborating with OpenAI to develop an AI assistant that facilitates communication with clients and offers personalised content
• Deutsche Bank is mulling utilising an avatar to assist employees in navigating internal systems and address HR-related queries
• Capital One has leveraged LLMs to enhance its customer-facing applications, providing tailored recommendations and addressing customer queries
• Provinzial Konzern is using generative AI to create synthetic data, expanding its customer data for training recommendation systems
• Wells Fargo has incorporated LLMs into its operations to improve customer service and support through AI-powered virtual assistants and chatbots
• Bank of America's AI virtual assistant Erica has gained popularity with millions of active users and is seamlessly integrated across various banking services
• Zurich Insurance is exploring the potential of LLMs for data extraction and analysis from claims descriptions, as well as code generation in statistical models

This is just the beginning. We foresee the adoption of generative AI in the BFSI industry running over years, if not decades.

The key impediments for now are unavailability of high-quality data and inability to integrate generative AI into existing processes.

Further, the new technology comes with significant risks. Some major banks have restricted the use of ChatGPT in response to concerns about sharing sensitive information.

Striking the right balance between automation and human involvement is essential for maintaining customer trust and mitigating potential risks.

We expect financial institutions to traverse this journey cautiously, albeit optimistically.

References - 

• Wells Fargo’s New Virtual Assistant to Be Powered by Google Cloud AI : October 24, 2022
• Morgan Stanley Wealth Management Announces Key Milestone in Innovation Journey with OpenAI: Mar 14, 2023
• JPMorgan is developing a ChatGPT-like A.I. service that gives investment advice: May 25, 2023
• How We're Using AI and ML : By Capitolone.com/machinelearning