Faster time to market — use cases deployed in 2 months
Use cases to be delivered over the coming year
Empowered business and IT users
Over the years, ABN AMRO has grown to become the third-largest bank in the Netherlands, generating a wealth of data with hundreds of terabytes of data across hundreds of disparate data sources — but their approach to data and analytics was far from where they needed it to be in order to realize their data-driven aspirations. Hampered by a centralized on-premises data warehouse and inefficient workflows, they struggled to transform into a truly digital bank that could realize the full potential of their data.
“In order to capitalize on the opportunity in front of us, we needed the ability to harness the data at our disposal and work in a different way than we are used to,” explained Stefan Groot, the head of analytics engineering at ABN AMRO. “Our organization suffered from highly disparate teams working in isolation without true collaboration and sharing of best practices. Although people were communicating, they weren’t working together around data as a cohesive organization, which created mass inefficiencies.”
Groot understood the importance of moving away from their traditional big data warehouse toward a domain-driven lakehouse and a data mesh approach where the creation and processing of data is federated across domains by using a proven, standardized platform.
He also envisioned an InnerSource culture among the hundreds of data professionals in which everyone is working collaboratively and sharing data insights seamlessly to support business goals.
With the goal of transforming the bank into a data-driven organization, ABN AMRO shifted toward the Azure cloud for ease of use and scalability and was able to operationalize the Databricks Lakehouse Platform within a couple months. With Databricks, they are now able to accelerate innovation by introducing data-driven solutions that improve strategic decision-making, operational efficiency, cybersecurity and the overall customer experience.
Databricks eases access to multiple data sources and simplifies infrastructure management at any scale. Delta Lake allows ABN AMRO to create data pipelines that are not only fast but highly reliable — critical for analytics and the data science teams who rely on complete and accurate data for decision-making, analytics and model training.
With data at their fingertips, ABN AMRO’s data analysts are now able to easily perform analytics and transform data to feed business reports and dashboards via PowerBI.
The data scientists and machine learning engineers can also easily deploy models into production with MLflow. What used to be a highly disjointed and manual process filled with inefficiencies is now consistent and automated.
With Databricks, including a new marketing automation stack, ABN AMRO is now able to react to customer preferences faster, allowing them to serve relevant product recommendations or deliver a relevant service to ensure satisfaction and reduce churn.
Fraud detection is also a focal point for the bank. Through the use of machine learning, they can identify anomalous behavior to prevent fraudulent activity such as money laundering, which is a common means to fund terrorism around the world.
Business operations is also an area that is greatly benefiting from data and AI. From a customer support perspective, Databricks powers a client dashboard that provides a complete view of the customer and all their assets and transactions in near real-time. This provides the insights their help desk and support teams need in an easy-to-consume format to better serve their customers. Additionally, they can monitor how many calls each support representative has made and identify opportunities to improve efficiencies and reduce operational costs.
Today, ABN AMRO has flipped their data strategy on its head. Rather than technology being the problem, now it’s an enabler. Because of this paradigm shift, the data teams at ABN AMRO have been able to deliver nearly dozens of use cases across the business at lightning speeds — 10x faster than with their existing infrastructure. But that’s just the tip of the iceberg. Groot estimates that they’ve only delivered 5% of the total use cases they have on their road map, with over 100 additional use cases, and even more machine learning models, planned over the coming months.
Most importantly, as the company grows so does the data team’s ability to grow and support the business. “We are a truly agile organization now,” expressed Stefan Boom, Solutions Architect at ABN AMRO. “There were things that weren’t even possible before. Now we have the flexibility and capability to do anything.”
With Databricks playing a pivotal role in their data-driven decision-making, ABN AMRO has opened the floodgates to empower over 500 team members across data engineering, analytics, data science and the business to leverage data to experiment and build solutions that deliver on use cases across the organization.
Looking ahead, Groot feels that in the coming year they will see exponential growth in models with 100+ additional use cases on the road map. Now that they have the technology stack, organizational structure and processes, and collaborative culture in place, they will be able to shift their focus and accelerate value creation across the enterprise.