Case Study

Elevating Data and Analytics Maturity Index and improving Data Governance through Modern Analytics Platform

Catalysing an analytics-led transformation for one of Europe’s largest credit management companies to maximize return on investment on their data assets

About The Customer

The company was formed in 2015 after merger of the UK and German market leaders. In 2018, they completed another acquisition in the Nordics region. With a core mission to make credit work better for all, they leveraged combination of data analytics insights and robust risk management to provide expert solutions in debt purchasing, third party collections and business process outsourcing. They wanted to improve time to market for their data products with a multi-country strategy for data consolidation, focus on minimizing cost of data management with zero risk to data security and enable analysts and data scientists with easy data discovery and data product development.

Business Challenge

With 8 lines of business across 9 different countries in the EU, the client inherited 500+ source systems that were regularly accessed by 3000+ users. They wanted to increase the speed of bringing highly accurate and reliable predictive AI/ML models to market. In a growing industry with increasing credit risk exposures and frauds, it was critical for the client to foresee and minimize these potential risks through effective data models in scale.

Following are the challenges faced by the client:

  1. Isolated and Disparate systems/data sources
  2. Extremely high time (8+ months) to build analytical models for due to manual work, lack of data preparation etc.
  3. Lack of tooling for self-service BI
  4. Cross department alignment on data processes
  5. Maintaining regulatory compliance
  6. Lack of roles and accountability for data management
  7. Inefficiencies and costs for data management
  8. Report rationalization
  9. Lack of advanced analytics for data monetization

Approach to Modernization of Data and Analytics Platform

Brillio provided a multi-year roadmap for data transformation on cloud powered by BrillioOne flash.ai. The platform brings multiple set of accelerators to building SCALABLE HIGH ROI ANALYTICS PLATFORMS  by enabling customers to accelerate time from concept to implementation, improve trust on data / analytics, provide better user experience leveraging the data engineering, advanced analytics, and Insights as a service on Azure. The solution was rolled out in a phase-wise manner as follows:

  • A 4-week design-thinking led assessment of data landscape with current operations and business aspirations, executed through ai DASH
  • Formulation of a data strategy for robust platform development
  • Development of advanced analytics platform leveraging ai ML workbench
  • Implementation of data management layer to maximize platform usability and RoI
  • Report rationalization through ai CRED and BI & Self-service enablement for Insight democratization
  • Enablement of data scientist toolkit

Business Impact and Benefits

Brillio’s assessment led approach enabled by the Flash.ai DASH highlighted the client’s data and analytics maturity across 108 sub-dimensions with respect to the industry average. The resulting roadmap addressed current gaps with more appropriate technological drivers that can help the client progress higher in the maturity index with efficient governance.

By mapping the data strategy with user needs ranging from citizen data scientist to IT expert, we created a single source of truth for data standardization of processes, aggregation of KPIs and inculcated strong data governance mechanism.

Building the advanced analytics platform with reduced time to market was critical to a successful transformation undertaking. With Brillio’s Flash.ai platform, we leveraged accelerators that reduced onboarding of new data sources by almost 50% such as

  • APTA, an automated testing framework, to comprehensively manage data quality for effective build and maintenance
  • Script centre and CLIP, to optimize the architecture to include high volumes of near-real time data
  • ML @ Scale and CRED to enable self-service data preparation capability for data science team
  • AI-driven data quality framework to self-learn and resolve quality issues

Powered by Flash.ai Trust Suite, we implemented a data management layer to ensure that PII data is protected and follows GDPR standards. The platform cohesively tackled all data management needs from data discovery, data lineage, data quality to security and compliance leading to 70% reduction in data preparation time.

The self-service business intelligence portal empowered internal and external users with standardized KPIs across regions and reduced organization inefficiencies with centralized development and distribution of reports. The portal was enabled with the features such as,

  • Real-time data quality monitoring and remediation,
  • ML-driven rules management engine powered by 500+ pre-defined rule for integration with data lake, and
  • dedicated data quality dashboard for monitoring and notification
  • Seamless integration with data stores and analytics applications

Using multitude of pre-built scripts hosted in Flash.ai’s Script Centre, we built custom products for data science enablement that made most used data sets available for access at anytime with a log of state of each variable over time. This resulted in 90% reduction in query execution time on a like-on-like basis for 12000 variables made available through the feature store cutting down model development time from 7 months to 2 weeks.

The multi-year comprehensive data and analytics modernization endeavour resulted in return of investment of ~3.7x with a profit upwards of $17 million.

A significant amount of time has passed since my last article on this topic. One of the exceptional qualities of […]

@Fahim Khan • Nov Tue 20

The workplace’s infrastructure and design are a crucial part of any business’s key decisions and are important for its growth. […]

@Siva Perubotla • Nov Tue 20
Latest Case Study
LinkedIn Instagram Facebook Twitter