Swedbank is a multinational bank with a 200-year history. Recently, the company took the bold decision to migrate its big data platform to the cloud in search of greater scalability and security. With Azure Databricks functioning at the core of this solution, the successful migration has brought the bank a wealth of benefits. From reduced time-to-market and cost savings to significant improvements in its ability to detect fraud – it has set the bank and its 7.7 million customers on course for a more secure future.
“Security has evolved in banking. Where it was once the domain of IT and cybersecurity teams, now it is everyone’s business. At Swedbank, we’re all involved in the fight against cybercrime.”
Rikke Berner Nilsson, Chief Product Owner Financial Crime Analytics, is describing a key aspect of the evolution of Swedbank, the company she has seen change and develop dramatically over her five years with the organization.
“A cultural change has taken place within the company. There used to be apprehension about the cloud, particularly around data security,” says Nilsson. “But that has given way to a mentality in which we have embraced what the cloud can offer us.”
That shift has been motivated in part by a need for greater scalability and flexibility as the bank seeks efficiency gains and faster time-to-market for cloud-based innovations. But also by a tightening of data retention rules and the need to provide peace of mind to customers across the Nordics and beyond.
“Ensuring we are in compliance with regulatory authorities is essential,” says Nilsson . “At the end of the day, we are a bank looking after yours, mine and everyone's money.
“Trust is incredibly important for us,” she adds. “Which is why we have a zero-tolerance approach when it comes to cybersecurity. The migration to the Azure Cloud has been a big part of realizing our vision in the fight against financial crime.”
A bank embracing a cloud-first approach
Headquartered in Stockholm and serving customers in Sweden, Estonia, Latvia and Lithuania, Swedbank is a market leader in mortgages, retail banking, asset management and other financial services. The bank serves over 7.7 million customers and employs around 16,500 staff in 216 branches.
“We’ve always been an inclusive bank, one for the many, and that spills over to the culture of working in Swedbank,” Nilsson says. “The hierarchy is very flat, and we choose the best person for the job when we have a challenge.
Ensuring it remains a dynamic and inclusive workplace, and continues to be at the forefront of banking in its home country and beyond has meant prioritizing digitalization. This change in priorities for the bank has also spurred a cultural shift, which is helping to drive its transformation as it embraces the cloud.
“The flexibility and scalability of the Microsoft Intelligence Data Platform, with its enhanced security, became a north star for us to steer a new course,” explains Ragavendra Singh Jaswal, Team Manager for Cloud Engineering, Group Business Intelligence at Swedbank. “One that would allow us to benefit from the AI and machine learning capabilities enabled by Azure Databricks, using automation to further enhance, accelerate and streamline our security measures.”
Proving the case for full migration with Microsoft
The team at Swedbank now needed to make the case for migration from their existing Big Data on-premises set up, based on Hadoop clusters along with a heterogeneous and complex ecosystems, to Microsoft Azure. Security was a key consideration.
“We started a proof of concept in 2020,” says Swedbank’s Solutions Architect Manish Singhal. “One part of that was the technicalities of the solution design. But the other focused on legal compliance, information security, and having risk assessments ironed out and agreed together with our committees.”
Once the team were happy that the setup was fully compliant and in line with the requirements of the relevant stakeholders, they set about building the new infrastructure.
“Data is ingested from various different sources and flows into the Azure Data Lake, where the data is cleansed and modeled. The Data Lake is also the primary service for hosting of all our data,” Singhal explains. “It’s then processed via Databricks, which is the core of all computation. We now employ Databricks for all our use cases, and all the stakeholders are using it to build their models.
“Everything revolves around the CI/CD process, so next Azure DevOps plays a major role in terms of migrating cases from development to production. In other words, Azure DevOps plays a vital role in terms of overall migration strategy.
“Finally, Azure Data Factory is used as a scheduler or orchestration tool to host or schedule these models in production set-up,” adds Singhal.
He is keen to emphasize the importance of Microsoft DevOps: “We spent a lot of time together with Microsoft while we were designing the whole DevOps architecture, because we essentially think of DevOps as glue – it stitches all your components together and ensures that there's proper CI/CD.”
Scalability with security cuts time-to-market, costs – and crime
Swedbank is already enjoying real-world benefits thanks to the migration, explains Jaswal. “The goal of the migration was to decrease time to market while increasing scalability, resilience and availability. With the Microsoft Cloud, we’ve achieved that.
“We can now onboard data into the platform from multiple sources and provision to different systems with reduced cycle times. Once the model is ready for production the whole CI/CD cycle has been automated in Azure DevOps,” he continues.
The new infrastructure has also brought significant improvements in the capabilities of their anti-financial crime unit, their AML models and scanning capabilities. This has enabled Swedbank’s teams to identify money laundering and other criminal activity more effectively. “Using AI and machine learning modules, we’ve been able to reduce the number of false positives,” says Nilsson.
And the time-to-market benefits are clear too. “It’s much faster for us to deliver a use case to production now,” says Singhal.
“And these benefits are more visible to our customers and stakeholders,” he adds. “Previously, it was tough to demonstrate. But deploying Azure, and working together with Microsoft, there's more visibility for our stakeholders.”
Educating an organization
With an established and highly successful use case under their belt, communication of these positive outcomes is at the center of the team’s strategy going forward. “We want to spread the benefits of this approach throughout the organization,” says Nilsen.
It’s a challenge that the entire team feels is worth the effort, according to Singhal. “It takes a lot of effort to turn a big ship like Swedbank around, but we’ve got the right crew, and more than enough motivation,” he enthuses.
As Jaswal puts it: “We want to grow together with the rest of organization, and ensure that there are more and more stakeholders onboard with this platform.”
“We’re on a mission,” he concludes. “To ensure our internal teams are managing their day-to-day work in the most efficient way possible. Our partnership with Microsoft is helping us to achieve that, driving us towards future targets.”
“Security has evolved in banking. Where it was once the domain of IT and cybersecurity teams, now it is everyone’s business. At Swedbank, we’re all involved in the fight against cybercrime.”
Rikke Berner Nilsson, Chief Product Owner Financial Crime Analytics, Swedbank
Follow Microsoft