Published on January 23, 2024, 6:06 am
Generative AI, a branch of Artificial Intelligence (AI), has found practical applications in the banking sector. According to Michael Abbott, Accenture’s global banking lead, generative AI is being utilized to generate summaries of call center interactions, provide employees with co-pilots, ensure compliance with loan rules, and develop new core systems.
The interest in generative AI among banks is on the rise. A survey conducted by KPMG revealed that 49% of financial services firms are implementing this technology, with 19% already reaping its benefits. However, questions remain about where to start, how to monetize it effectively, and how to prevent unintended consequences such as hallucinations or data privacy violations.
Abbott believes that generative AI has the potential to boost banks’ productivity by approximately 20% to 30%. The technology offers operational efficiencies that can streamline various banking processes. For instance, banks have implemented generative AI for post-call recordings in call centers. Instead of spending four to five minutes summarizing customer conversations manually, generative AI can perform this task within seconds. The human call center representative can then review and confirm the accuracy of the generated summary before incorporating it into records.
Another example involves mortgage loan origination providers using generative AI to analyze loan applications and compare them against Fannie Mae requirements. This process allows for a quicker identification of potential issues or red flags instead of manually reading through all the documents.
Despite these operational improvements, concerns arise regarding job losses due to increased automation. However, Abbott states that many banks view generative AI as a means to eliminate wasted time spent on non-value-added tasks rather than reducing headcount. By freeing up employees’ time from mundane activities like summarizing calls manually, they can focus more on engaging customers and maximizing cross-selling and upselling opportunities.
Banks also recognize the revenue opportunities presented by generative AI. For instance, it can help develop customized savings scripts for customer service representatives to determine the best interest rates to offer customers, optimizing deposit beta rates, and personalizing recommendations based on customer behavior. These revenue-focused applications demonstrate the potential profitability generative AI offers.
However, potential risks associated with generative AI cannot be ignored. Banks and companies must adopt a responsible and ethical approach when implementing these models. Abbott stresses the importance of using a human-in-the-loop approach, where generative AI augments human decision-making rather than fully automating it. This ensures that any risks such as bias or hallucinations are mitigated.
The implementation of generative AI in banking often follows a cautious, aggressive approach. Banks take their time to thoroughly understand the technology’s impact on customers and ensure all standards and requirements are met. However, they remain proactive in experimenting with and internally adopting generative AI models to explore their potential benefits.
Even banks with outdated core systems can benefit from generative AI. Through reverse engineering techniques, legacy COBOL code can be analyzed, modernized, and transformed into next-generation code within existing systems. While not perfect yet, this process significantly reduces the time and effort required for system upgrades without completely replacing the core infrastructure.
In conclusion, generative AI presents numerous practical applications in banking, allowing for increased operational efficiencies and revenue generation opportunities. The cautious adoption of this technology demonstrates banks’ commitment to ensuring customer safety while leveraging its potential benefits to enhance productivity and customer satisfaction.