ChatGPT’s release has revolutionized AI, and industries have created innovative ways to deploy this tool and other models to enhance customer communication. Most recently, AI use cases have emerged in the banking, financial services, and investment (BFSI) industries that extend far beyond language processing capabilities.
How are financial institutions pioneering various AI tools for data generation, predictive analytics, and more?
NVIDIA conducted a survey demonstrating key opportunities for AI’s application in financial services. The survey revealed significant growth in use cases for natural language processing (NLP), large language models (LLM), and recommendation systems.
Practical applications for NLP and LLM exist within the context of contextual banking, where organizations have utilized tools like ChatGPT to gather insights on individual transactions and maximize the customer experience. Similarly, capital markets have leveraged these language models to predict stock market trends to manage assets and optimize investment portfolios. The final use case observed for this AI form involves fraud detection and identity verification related to banking and transactions.
Recommendation systems generate and analyze data to help consumers make informed financial decisions. These engines also assist organizations and internal stakeholders in recommending resources to their consumers. Spiralem’s Managing Partner, Bruno Diniz, cites an example of recommendation systems, stating that they can provide “financial information on a [case-by-case] basis that really reads information on the client, understands the problem they have, and makes an explanation specifically for a situation that perhaps the client has not even been able to assess themself.”
Many financial companies face challenges generating conversions, customer satisfaction, and retention. Recommendation systems and generative AI models provide opportunities for personalized content. Case in point, Capital One wanted to optimize conversions through a relevant homepage banner placement. Kevin Levitt, NVIDIA’s leader of Global Business Development for Financial Services, describes Capital One’s efforts: “They leveraged deep learning and NVIDIA’s application framework for recommendation systems…to understand and identify the optimal placement to give any consumer after they’ve logged in.” As a result, Capital One increased conversion rates by 60%.
Another company based in Brazil has deployed a generative AI tool to create customized content for asset management, business proposals, and hyper-personalization to meet individual needs and banking preferences. You can also use these models to develop optimized marketing and advertising copy.
Before piloting any AI model, it’s crucial to test its potential value for an organization by conducting proof of concept. Two sectors of BFSI have begun deploying testing for these models: compliance as a service (CaaS) providers and capital markets.
One CaaS organization has integrated ChatGPT to help banks comply with complex data-sharing regulations existing within federal and state privacy laws. Independent fintech advisor Efi Pylarinou explains how this company has executed PoC: “They are training and getting meaningful answers in terms of the risk that the financial services organization is running due to the complexity of these [regulations]. This highlights how important it is in terms of what prompts you’re asking these models and how you’re training them.” This business has successfully determined the value of ChatGPT in risk assessment and management.
Fintech companies are training LLMs to digitize documentation such as earnings reports, providing valuable insights into investment share prices for capital markets. Yet LLMs like ChatGPT supply general, nonspecific insights, so experimenting with multiple models is paramount.
Multiple institutions in various sectors have demonstrated the emerging potential for AI in BFSI and the value it can deliver to consumers.
