UBS has launched an extensive data transformation initiative aimed at the development of two specialized AI assistants, referred to as UBS Red.
These innovative tools leverage Azure AI Search and Azure OpenAI Service to aid client advisors during their interactions with customers. Notably, UBS Red features multi-language capabilities, which enhances the efficiency of meeting clients’ diverse needs.
Michel Neuhaus, who oversees AI, data, and analytics for Personal & Corporate Banking at UBS, emphasizes the magnitude of the project: “The breadth of our internal information, financial acumen, and insights into global financial markets is immense. We need to ensure that this knowledge is readily accessible to everyone in the bank, and Microsoft assists us in achieving that.”
In response to this extensive information reservoir, UBS established a knowledge base that can be queried. By incorporating Azure AI Search, the bank developed enhanced retrieval techniques to derive new insights from unstructured data through language model reasoning. This improvement in vector search capabilities allows for the discovery of information based on conceptual relationships among diverse data formats, enabling UBS to maximize the value of its extensive data resources.
Claire Corish, head of analytics and sales platforms within UBS’s Operations and Technology Office, stated, “We have digitized around 60,000 documents related to investment advice and products. Now, our employees can easily access this queryable knowledge base built on Azure, substantially reducing the time required for meeting preparations and research.”
The accessibility of its data reservoir has empowered UBS to create an internal AI Hub dedicated to exploring new applications.
Looking ahead, Jonas Isliker, head of AI, data & analytics for Global Wealth Management at UBS, shares, “Our primary focus is to swiftly embrace AI transformation. While the rollout of key AI solutions like Red is vital, developing a scalable organization capable of accelerating AI project delivery is equally critical. This entails adopting a data mesh, transitioning to the Azure cloud, establishing robust AI risk governance, and launching AI factories for faster, more adaptable solutions. All these initiatives are essential to meet the increasing demand for AI applications throughout the organization.”
In 2022, UBS announced ambitious plans to migrate more than half of its applications, including essential workloads, to Microsoft Azure within five years. In the past ten months, the Azure OpenAI Service has been implemented across the bank’s primary divisions globally, reaching 30,000 employees by December 2024.