Swiss Re Drives Deeper, Faster Insights with Azure Synapse Analytics and Power BI

To deliver on these services, Swiss Re has developed a vast analytics and business intelligence (BI) infrastructure that serves multiple business units around the world. However, for its Property & Casualty Reinsurance division, the insights and reports from that system weren’t coming fast enough. Building the reports was a labor-intensive process that required technical staff to navigate the hard-coded system. Report updates were also bundled into a single release cycle, which meant that if a new request did not make it into one cycle, it would have to wait until the next one began. “These cycles only occurred four times a year. Eventually, we got to a monthly cycle but only when there were no changes to the data model,” notes Miguel Marin, Head of Business Intelligence, for Property & Casualty Reinsurance. The team needed a single source of truth and a self-service solution to help keep up with their analytics. 

To meet these needs, Marin and his team built an in-house solution called Cockpit, a centralized BI reporting solution that helps centralize data into a one-stop-shop.  While the system met the initial objectives of providing basic data visualizations and drilldowns, it was starting to show age and couldn’t deliver data as quickly as the company needed it. After migrating about 650 reports into Cockpit, Marin quickly realized that Cockpit was unable to handle the size of loads that Swiss Re needed. There were also functionalities and capabilities Marin and his team needed but would be too time consuming and costly to build in-house. The team decided to invest in a solution that would handle the large amount of data in an automated way that would also reduce processing time.  “We all realized that while this seven-year-old solution had served its purpose, our team needed a faster, more advanced set of data intelligence tools” says Marin.

Reimagining data intelligence

Around this time, the Swiss Re P&C Reinsurance team came across a Gartner report on business intelligence which ranked Microsoft the leader in business intelligence with Power BI. Since Swiss Re had a strong partnership with Microsoft, Marin and his team began discussing the possibility of using Power BI as their main reporting tool. By working with Microsoft, Swiss Re had the knowledge and foundation to execute a fact-based PoC to determine if Azure and Power BI were a good fit. Before officially deciding, they needed to determine if Power BI would fit with Cockpit and the rest of their preexisting tools and expanded the PoC to four more use cases. After seeing a successful compatibility with their current tools, Swiss Re quickly realized that they could create an even stronger solution by combining Power BI with Azure Synapse Analytics. 

Swiss Re was impressed by several Azure Synapse features, such as the flexibility of Azure Synapse to scale services up and down during peaks or loading times. With Synapse, the load chain was reduced by 5 hours. Cost savings were another stand-out factor. When evaluating cost, Marin and his team looked at several requirements, including consumption, services, and components. “We came to the conclusion that if we use the components in Azure, we could cut our operational costs dramatically, because we only pay for what we use,” Marin points out. “Plus, by reserving Azure services in advance, we have saved 30 to 40 percent on several services.”

In theory, moving to a managed, cloud-based infrastructure like Azure Synapse would address many of the key challenges that Swiss Re faced with its on-premises system. Marin and his team wanted to confirm that an Azure system would not only deliver all the functionality of the existing system but also free Swiss Re from its dependency on the data marts used to retrieve data. 

To ensure that Synapse would realistically align with their needs, the team set up a proof of concept. The review process included a long list of security requirements needed for internal and regulatory compliance, and approval through the company’s rigorous data governance framework process to expose property and casualty reinsurance data in Azure. Swiss Re evaluated several competitors, but in the end Azure Synapse and Power BI came out on top. 

“With Azure and Power BI, we're not dependent on another internal system. Deploying new data or adding a new component was now just a few clicks away” explains Marin. 

Bringing the proof of concept to life

With a clear understanding of how Marin’s team wanted the data and analytics platform to look, they jumped into building out their solution. Power BI services were set up and put into production allowing for a single source of truth for all Property & Casualty reinsurance data and keeps all data in sync across business applications and reports. 

Data was then migrated from the existing databases to Azure, where data marts were built on Azure Synapse. In the current solution, Azure Data Factory and Azure Data Lake Storage bring together structured, unstructured, and semi-structured data such as logs, files, and media. Azure Databricks is then used to clean and transform structureless data sets and combine them with structured data from operational databases or data warehouses. Azure Databricks also enables scalable machine learning, deep learning techniques, and the ability to perform root cause determination and raw data analysis.

Native connectors integrate Azure Databricks and Azure Synapse Analytics to access and move data at scale. With the help of Azure Synapse Analytics and Azure Analysis Services, the data warehousing and compute environment are supported by a massive parallel processing architecture. The large parallel processing architecture allows data to be processed quickly and efficiently. 

After all the data is processed and deployed, it is pushed into numerous Power BI Premium data sets. This enables better performance, and once data is loaded the system can be paused to reduce costs. The deployment also covers the development of a data governance framework. A new Cockpit user interface, with improved look and feel, along with new capabilities, like machine learning and data engineering, provides a simple and visually pleasing experience for end users.  

Creating a more efficient reporting system

The P&C Reinsurance Division’s old report-building structure consisted of approximately 150 reports and screens in Cockpit and 100 MSBI and MSBQ reports from an Excel-based connected database. Data within the Property & Casualty Reinsurance Division at Swiss Re is used for 3 different types of analytics: operational reporting, operational analytics, and investigative analytics. 

Operational reporting provides data (typically to dashboards) in a predefined format with standardized metrics to provide more insight into business processes. Many operational reporting dashboards are focused on managing claims and contracts and helping users leverage data to quickly make decisions. These dashboards typically utilize Power BI Embedded so they can easily be added into a user’s normal workflow. By prioritizing design and user-centric reports, Marin and his team could leverage Power BI to build solutions for a specific target audience by capturing their specific data needs and what role analytics plays in their decision making. 

Operational analytics involves more ad-hoc reporting, striving to enable new combinations of established operational data to meet unique user needs. Users leveraging operational analytics are typically pursuing data investigations or prototyping to gain unique metrics. Some operational analytics dashboards also use the Q&A functionality to help users get tailored insights, in addition to leveraging Azure Maps for geospatial analysis.

Investigative analytics is the most complex reporting format for Property & Casualty reinsurance data. These reports typically require advanced analytical techniques such as AI, data mining, and predictive analytics to help understand complex data, find patterns, or do a deep dive analysis. 

Seeing the benefits of a powerful partnership

Swiss Re has reaped tremendous benefits since integrating Power BI. “With Power BI, we are enabling our business users to perform their own advanced analytics cases in ways that simply weren’t possible with our hard-coded system” says Marin. The team has also seen better visualization possibilities after working with their design team to set design standards and maintain the best UX practices. The sheer speed of development has also impressed Marin and his team while allowing them to figure out what is working and what isn’t. “We can develop a dashboard for a whole unit in three to five days, compared to months with the old process. That is just an unbelievable jump in efficiency.”

The new Azure environment also delivers faster loading times—almost 40 percent faster than loading the same amount of data with the on-premises system. Now, the load chain has been reduced by 5 hours. “This not only saves us time per loading process but enables us to load fresh data more often” notes Marin. The new environment is now handling a huge amount of data, which currently totals over 250 base tables and more than 600 million rows. Swiss Re has also been able to eliminate extra costs with Azure Synapses’ scalability and ability to easily avoid running extra processes. 

Word of the P&C Reinsurance division’s success with Power BI has spread fast and has been adopted across adjacent teams. Marin attributes the fast adoption to client-centric design that makes dashboards easy to understand and gain insights from. Additionally, the model used to deploy the technology within Property & Casualty Reinsurance has paved the road for other teams. “With all the security and compliance approvals completed, other divisions can deploy Azure Synapse even faster using our model,” says Marin.

Because of the security and governance infrastructure Marin and his team have built in collaboration with Microsoft, other teams and divisions can accelerate their cloud migration and get Power BI capabilities and improve reporting practices even quicker. It all adds up to faster, more efficient analysis and deeper insights. For Swiss Re, that might just be the most important business advantage of all.