GXS Bank Builds a Strong Data Culture for AI‑Driven Decision-Making

Challenge: Lay a data-driven foundation for trusted AI

Dr. Geraldine Wong is the Chief Data Officer at Singapore’s all-digital bank, GXS; she’s banking on AI.

GXS, a banking consortium backed by Grab and Singtel, launched in 2022 to serve a potential market of three million customers, with a particular focus on consumers, entrepreneurs, and the self-employed. GXS offers novel banking solutions such as their GXS FlexiLoan, which combines features of a credit line and a personal installment loan to give consumers more control and flexibility in managing their finances.

Dr. Wong believes that AI can be a strategic differentiator for the bank. “We're looking at leveraging data assets, especially the ecosystem data of Grab and Singtel, to better our customer experience, offer better products, and improve risk management using data,” she says.

Dr. Wong knows that trusted AI results need trusted data. “Definitions need to be accurate in order for the large language models (LLMs) to use the data,” she says. “If you're fine-tuning your models or pretraining some of these models, it's of utmost importance to have high data quality.” She also believes clear data ownership is critical to ensuring only trusted data is used in GenAI models.

Before the bank even launched, Dr. Wong and her team set out to build a strong data culture at GXS. “In traditional settings, decisions get made based on experience and gut feel,” she says. “In an ideal state, data culture would be everyone from the board to the CEO, from finance to HR, breathing and living with data. The first thing they ask is, ‘What does my data tell me?’” To reach that state, Dr. Wong’s team needed to ensure that data used for decision-making, with or without GenAI, was discoverable and trustworthy.

Objective: Build a strong data culture to support AI-driven decision-making

To create a strong data culture that supports the use of GenAI for strategic decision-making, GXS needed to:  

• Create a single source of truth for data 

• Define data terminology and ownership

• Implement data governance to increase trust

Implementation: Creating a single source of trusted data

GXS chose Alation as an integral component of their cloud-native infrastructure, which includes Snowflake, AWS, and Tableau. They chose Alation because of its user-friendly interface that provides value to both technical and business users.

Alation is the foundation for data governance at GXS. As a new bank, GXS used Alation for metadata cataloging and tagging as they created their Snowflake data lake. As new data is added to Snowflake, it is tagged in Alation to guide users in the proper use of that data, including consumers’ personally identifiable information (PII). They also used Alation to identify data owners and to create a glossary to clearly define terminology. GXS relies on Alation’s lineage capabilities to trace back source tables and understand downstream impact as part of enabling users to better understand and discover their data assets.

 “Alation provides the platform and capability for stakeholders — business stakeholders as well as enterprise — to be able to search for the data they need, and for them to understand that the data that they need is defined such that they have confidence in using that data for their business needs,” says Dr. Wong.

Results: Trusted data drives trusted AI

Dr. Wong and her team have built a strong data culture at GXS; nearly all of the bank’s 300 employees can access Alation to find trusted data. In the past year, users have executed over 3,000 SQL queries and conducted over 5,000 searches using Alation’s Google-like user interface. Today, Alation provides the single source of truth for GenAI modeling and data-driven decision-making at GXS.

“We’ve created a data culture where people know and trust the data they’re using,” says Dr. Wong. “They know where to find things, they no longer come to us for definitions, and they know the workflows and who the data owners are.” She doesn’t want to stop there. The next step is for people throughout the bank to instinctively know which data and AI capabilities to leverage for their use cases, whether they’re working on credit scoring, anti-fraud, or anti-money laundering. “It’s all about leveraging AI capabilities, data, and technology to drive these use cases and provide more convenience to our customers as well.”

“There's a lot of skepticism on what AI can do,” concludes Dr. Wong. “We need to trust the data that goes into the AI [models]. If organizations and their customers are able to trust the data that the organization is using for such models, then I think that's a good starting point to building that trust for AI governance or responsible AI.”