How financial institutions can leverage AI to transform data into gold
Data management remains a pivotal challenge for financial institutions, encompassing a myriad of complexities.
A panel of industry experts from Nomura, BNY, and Xceptor delved into these issues at the recent AFME OPTIC conference in London. They discussed the transformative potential of financial data, likening it to “gold” when optimally utilized.
According to Corlytics, despite the vast amounts of data accessible today, significant hurdles exist with granular, unstructured, or fragmented data scattered across various internal silos. As AI-driven systems gain prominence in trading and decision-making, incorporating these overlooked data sets becomes increasingly crucial.
Ling Ling Lo, global head of data strategy and transformation and EMEA chief data officer at Nomura, highlighted the importance of dark data—data that is unstructured, underutilized, and often ignored.
This type of data, while challenging to access and process, is key to developing next-generation trading and AI models. Traditional data, such as trading and client information, is readily available due to regulatory requirements. In contrast, dark data remains largely untapped due to its unstructured nature, found in documents and communications, which poses significant challenges for AI-driven analysis.
To fully leverage AI in financial services, institutions must shift focus from structured to more complex, unstructured data sources. This approach could dramatically enhance AI-driven processes, giving banks a significant competitive advantage through more sophisticated predictive algorithms.
Archie Jones, SVP of data and AI ethics at BNY, underscored the importance of AI and machine learning in enhancing commercial outcomes through refined data usage. The potential of generative AI to convert unstructured data into actionable insights could revolutionize capital markets, where granularity and real-time analysis are invaluable.
The role of AI in automating data processes was also discussed. AI’s capability to perform sentiment analysis on communications helps firms extract valuable insights from everyday interactions. Generative AI is particularly notable for its efficiency in summarizing and processing complex documents, which could lead to significant time and cost savings.
However, this advancement comes with risks, such as AI-generated inaccuracies, or “hallucinations,” due to poor data foundations. Dan Reid, chief technology officer at Xceptor, emphasized the importance of Retrieval Augmented Generation (RAG), which allows AI models to pull data from specific, proprietary documents, ensuring more reliable outcomes.
The transformative shift towards cloud adoption in financial services marks a significant development. The move to cloud platforms has been revolutionary, particularly in addressing legacy data challenges. Cloud solutions provide the scalability and flexibility required to optimize data models and transform legacy data from a cost burden into a valuable asset.
As pointed out by Xceptor’s CTO, it’s crucial that engineers have swift access to data, which cloud technologies facilitate, significantly speeding up data retrieval and utilization processes.
Trust in data and AI represents the next frontier in financial services. Ling Ling Lo stressed the necessity for firms to collect high-quality data to effectively build their models and utilize large language models to transform unstructured data into structured formats. As we venture deeper into the AI era, mastering data management in capital markets remains a crucial, ongoing journey.
