UOB and Standard Chartered Take a Governed Approach to AI
In a bid to enhance banking operations and eliminate inefficiencies, United Overseas Bank (UOB) and Standard Chartered are embracing artificial intelligence (AI) while navigating the complexities of regulatory constraints. Both banks recognize the potential of AI not just as a tool for innovation, but as a necessity for extracting value from their data.
Addressing Data Silos
UOB’s journey towards data transformation began in 2015 with the launch of its Enterprise Data Architecture and Governance (EDAG) program, utilizing Cloudera as its foundation. Initially, the bank faced numerous challenges, such as fragmented data sources, a lack of coherent data strategy, and low data literacy among employees. According to Alvin Eng, Head of Enterprise AI at UOB, the bank generated vast amounts of data across various business units, but much of it remained siloed within legacy systems, complicating cross-system analysis and slowing decision-making.
Moreover, UOB struggled with scalability. The exponential growth in data from digital banking transactions rendered legacy platforms slow and costly, hindering the adoption of advanced technologies like AI. Eng emphasized that “rigid data structures further slowed innovation,” making it unsustainable to continue investing in outdated infrastructure.
Standard Chartered’s Centralized Governance
Meanwhile, Standard Chartered sought a platform that could centralize governance without sacrificing the stringent regulatory and security requirements associated with on-premises data management. David R. Hardoon, Global Head of AI Enablement at Standard Chartered, noted that having governance controls such as native data lineage and policy enforcement was essential for ensuring high data governance and AI safety.
Preparing for AI Deployment
Before deploying generative AI use cases, Standard Chartered faced two main challenges: ensuring the availability of AI-ready data for reliable ingestion and maintaining access to adequate computing resources. Hardoon stated that collaboration with Cloudera helped address these challenges, allowing internal AI teams to efficiently route requests to designated generative AI models and conduct controlled experiments.
Conversely, UOB focused on overcoming issues related to data accessibility, quality, and scalability. The partnership with Cloudera enabled UOB to store, clean, and retrieve data more efficiently, addressing challenges like metadata consistency and data quality remediation. Eng explained that they established a robust data foundation supporting various AI initiatives, including more advanced use cases such as generative AI.
Innovative Solutions and Employee Training
UOB also introduced “playpens,” or sandbox environments, where teams could prototype solutions while adhering to governance standards. Eng noted that this approach encouraged experimentation among employees. To further empower staff, UOB organized multiple training programs and workshops aimed at equipping them with the necessary skills to leverage new technologies effectively.
According to Remus Lim, Senior Vice President, Asia Pacific & Japan at Cloudera, banks are primarily concerned with data unification, governance, and the responsible deployment of AI at scale. The objective is to generate real-time insights that enhance decision-making and improve customer experiences.
Business Transformation Achievements
UOB’s collaboration with Cloudera has led to the successful scaling of AI initiatives into production. These include AI models for portfolio optimization and customer engagement, now fully integrated into the bank’s operations. UOB has reported significant improvements in operational efficiency, particularly in cash management.
Previously reliant on manual forecasting for ATM replenishment, which was often inefficient, UOB’s data scientists built a predictive model based on ATM transaction data. This optimization resulted in a 33% reduction in replenishment trips, saving SG$1.3 million annually.
In the realm of risk management, UOB has developed behavioral indicators and machine learning models to identify unusual activities, thereby enhancing efficiency and accuracy in risk monitoring while reinforcing compliance and governance standards.
Standard Chartered’s Strategic Enhancements
Standard Chartered has also seen benefits, including tighter controls over personalization, improved precision in customer interactions, and more flexible management of data quality and compliance. Hardoon highlighted that the bank can now limit training datasets for model training to particular document types, bolstering risk oversight and AI safety practices.
Conclusion
Following the adoption of Cloudera’s platform, both banks have effectively managed, secured, and analyzed data across on-premises and cloud environments using a single governed architecture. This robust foundation enables UOB and Standard Chartered to operationalize AI use cases while complying with stringent regulatory standards.
As they look to the future, both banks remain committed to leveraging AI responsibly and effectively in their operations.
