Ageas.
How machine learning boosts performance, while reducing costs for insurance company

Case Study
Background

Every day, millions of people around the world depend on Ageas to protect their cars, their homes, their businesses and livelihoods.  

As one of the world’s largest insurance companies, it processes as many 5,000 requests for insurance quotes and approvals for underwriting, per second.  

Each of these requests must be screened by increasingly complex algorithms to check validity, while identifying potential cases of insurance fraud.   

However, this needs to happen quickly – within seconds. Failure to respond quickly to quotes in a tight timeframe increases the likelihood of the client purchasing elsewhere.  

Machine Learning for Insurer's Performance

The Challenge

Ageas runs this process via an on-premises licensed Hadoop cluster from an industry leading partner, which enabled them to go from a six week lead time to sub-second insurance quotes. 

However, the platform’s rising licensing and infrastructure costs were becoming prohibitive, while its quote performance was decreasing. The big data demands of the Ageas team could not be met by the existing solution or deliver on their goals to achieve real time fraud detection and adaptive pricing based on machine learning driven insights. 

With an underlying data set growing exponentially and algorithms increasing in sophistication, the team at Ageas recognised it was time for a change.  

Working to a tight timeframe of three to six months, the company wanted to perform a Proof Of Concept (PoC) with Microsoft, the company’s strategic cloud vendor.  

Moreover, in that same period, Ageas was keen to take this new solution to production and demise the legacy on premises cluster. 

The engineering team turned to data experts at The Data Analysis Bureau (T-DAB) to ensure they met their goals.  

The Solution

Ageas engaged T-DAB to provide strategic insight, architectural and development support for the new platform. 

A series of workshops with key stakeholders representing business teams, IT leadership and third-party security consultants, ensured the project got off to a smooth start and embedded best practices. 

Once the approved technology set was agreed with internal security teams, T-DAB took the lead in designing and developing an intelligent solution which incorporates the very latest cloud-native big-data tools from Azure.  

By drawing on Kappa architecture, T-DAB successfully incorporated a range of tools into the solution. Tools such as Azure Event Hubs, Azure Databricks, Azure Data Lake Storage, Functions and CosmosDB, were combined with a more traditional ETL orchestration layer (Azure Data Factory), to deliver a solution meeting the needs of the business, as well as its IT teams.   

From there, T-DAB worked sidebyside with Ageas’s engineering and BI teams to build an initial proof of conceptas well as a productionready implementation. 

The Result

In just three months, Ageas met its goals, thanks to support from T-DAB.  

Charged with designing and delivering a proven cloud-native big data platform on Azure, T-DAB’s close working partnership with the client meant all requirements from business teams and IT leadership were met, as well as those specified by internal and external security teams.  

The speed of the implementation allowed Ageas to accelerate plans for decommissioning the legacy on-premises solution, resulting in tens of thousands of pounds of operational cost savings. 

About The Data Analysis Bureau

The Data Analysis Bureau is data science and data engineering innovation company. We develop innovative, bespoke machine learning-driven solutions to allow anyone to infuse technology with the spark of predictive intelligence. 

to Find out more about the project & our services,
get in touch with the team.