Skip to main content

Redefining banking through customer-centricity

CLOUD: Azure

“Databricks was the platform I’d imagined from the start. By allowing us to manage data from across the bank, customers and partners, it has enabled Illimity to become the data-driven, innovative bank we set out to be.”

— Renzo Rognoni, Data Analytics Manager, Illimity

Banking is increasingly moving online, with many financial service providers now offering a digital-only experience. Illimity combines banking expertise with cutting-edge banking technologies to offer innovative digital direct banking services to retail and corporate customers. The performance of these technologies depends on the intelligence derived from the wealth of data that passes through the bank. However, their legacy architecture was built on disparate data warehouses, and Synapse lacked the flexibility they needed to deliver on their mission. It was vital that they migrate to a more robust and flexible data architecture before the bank launched — a feat only made possible by a partnership with Databricks. With the Databricks Data Intelligence Platform, Illimity is now able to process data in real time, extract actionable insights and securely share it with their strategic partners.

Legacy technology and vendor lock-in hinder ability to operationalize vast amounts of data

Illimity was always envisioned as an innovative digital bank, whose offerings and services would be informed by a wealth of wide-ranging data — from customer demographics to transactional activities. Prior to their launch, Renzo Rognoni, Data Analytics Manager at Illimity, was tasked with designing  the data architecture for the bank and the wider Illimity Group. But legacy constraints and lock-in from vendors meant that this wasn’t always easy.

With a huge amount of data to manage — both structured (e.g., bank account details) and unstructured (e.g., information from financial documents) — and a range of disjointed data sources, the bank wasn’t able to operationalize their data at the rate and scale required. The main source of data was Illimity’s core banking, from the accounting framework for servicing their customers’ transactions. Illimity also received and shared customer and transactional data with their partners that supported this core banking, such as payment broker Nexi and PSD2 gateway Fabrick.

“I wanted to create a lakehouse, where we could process and analyze the data in real time,” says Rognoni. “At that time, though, every data platform was far too rigid and regimented in its approach and meant that we’d be locked into a solution that simply didn’t meet our needs.”

Databricks took a different approach. They offered the flexibility and performance that Rognoni sought, for what he soon realized was a vision of the Databricks Platform.

Databricks serves as the foundation for the future of banking

Illimity’s data architecture is entirely built around the Databricks Data Intelligence Platform. Rognoni created a single source of truth, providing a unified view of all their data for downstream analytics and sharing with partners. With the lakehouse architecture as the underlying data infrastructure, Illimity chose Databricks SQL for data warehousing over other competitors like Synapse. This decision was further justified after Databricks SQL performed queries faster in a side-by-side comparison.

Rognoni cites the main reasons for using a cloud-based solution — Databricks Azure — particularly during this execution phase. “The most important factor was time to market. We had less than a year to get everything up and running. The Databricks managed solution allowed us to deliver everything in a really short space of time,” he says. “And, with applications created from scratch, we were able to anticipate any opportunity to optimize costs and deliver more value to our customers. What’s more, Azure is very capable of supporting the kind of strict constraints imposed on Illimity by regulators such as the Bank of Italy.”

Some regulations require processes that just don’t make sense in the cloud. In Italy, for example, one guarantee process requires banks to track the information of anyone who has access to their products. “Part of the required information is machine name and address,” says Rognoni, “For a platform-as-a-service, this is almost impossible to provide. Fortunately, with the combined support of Databricks and Microsoft, we were able to address this.”

The inclusion of Unity Catalog provides visibility over the lineage of all the company’s data and, with the lakehouse architecture, Rognoni’s team can now easily and securely unlock the intelligence from that data and extract insights to help the bank and their partners make informed decisions regarding customers’ accounts, automate transactional processes such as loan processing and mitigate fraudulent activities.

Creating a more accessible, personalized and ethical banking experience

Databricks has been a close partner of Illimity almost from the beginning. Over time, the nature of that partnership has evolved. From execution and performance to innovation, Databricks has always enabled Illimity to manage their data and intelligence, regardless of the solution. And it continues to evolve.

Databricks has supported Illimity’s adoption of LLMs to inform their knowledge base and power intelligent customer-centric chatbots. LLMs require vector embedding — a type of data representation that carries semantic information, which helps AI systems better understand that data. “Although we looked into several different solutions for vector embedding,” says Rognoni, “none of them were suitable. But Databricks had already done a lot of work on a framework that used the same technologies we were looking for, such as vector databases. It’s really on the edge of innovation.”

As a rapidly growing digital bank, Illimity has quickly established valuable partnerships with several technology players to sell their products to commercial customers. As part of this, it is important to securely and seamlessly share parts of their IP and components of their platform without sharing any of the data provided by their customers. To this end, Rognoni is now looking at how Databricks Clean Rooms can help shape the solution for monetizing Illimity’s architecture, using deterministic or machine learning models to share their code without compromising their confidentiality.

Rognoni considers Databricks a true partner today and into Illimity’s future. “While we had the imagination and ability to create a bank from scratch,” he says, “right from the start, we had to demand a lot from Databricks. And as we grow, they’ve continued to deliver.”