By Icon Solutions ·
The challenge
The client is the UK’s largest retail bank, serving over 30 million customers, and covers Retail Banking, Commercial Banking, and Insurance, Pensions & Investments. Their Economic Crime Prevention Platform (ECPP) provides technology services to the wider bank, supporting the bank’s strategies and policies against fraud and financial crime.
Icon was engaged to support a multi-year transformation programme to align the ECPP estate with a refreshed ECP strategy. This engagement required Icon to lead the definition of the platform target architecture and create product-level target architectures for the 30 in-scope ECP products.
The bank recognised that it needed to strengthen its capabilities in artificial intelligence (AI), machine learning (ML) and data as part of the programme. One key challenge the bank encountered was increasing the usage of AI & ML solutions across the platform, into areas such as customer risk assessment, customer profiling and alert management. A strong emphasis was placed on leveraging the bank’s in-house data science capabilities to reduce dependence on third party vendors.
With an increased focus on deploying more technologies within the bank’s digital framework, the need to enhance fraud prevention and anti-money laundering (AML) capabilities became vital. Increasing detection rates and reducing the number of false positives were key priorities in maintaining the bank’s position as a leader in the ECP domain.
Specifically, the bank needed to streamline its ECPP application estate and reduce the total cost of ownership of technical deployments. The elimination of data silos and duplication was imperative for the bank to reduce errors and ensure solution resilience.
It was also critical that the target architecture aligned with the bank's technology strategy, placing importance on adopting services-based and event-driven architectures, reducing technical debt and promoting cloud adoption.
The approach
The first step involved defining an ECP data reference architecture and its integration with other areas of the bank. This laid the foundation for defining the target architecture for each ECP product.
Key stakeholders were consulted, including domain/enterprise architects, data architects, business architects, product owners, engineering leads and senior business leads. A considerable amount of effort was spent discovering and understanding the bank’s existing reference architecture for data and the cloud, with Icon defining how the key principles would apply to ECPP.
Several inputs were used to shape the target architecture, the most important of which were:
- The bank’s data technology strategy
- The ECPP technology & data strategy
- The bank’s cloud reference architecture
- ECPP product canvases – which laid out the respective product’s scope, vision and alignment to ECP business strategy
- Icon’s Anti-Financial Crime reference architecture
Solutions and benefits
Icon consultants delivered a set of architectural solutions, including a product target architecture for the ECP data platform and the AI, Modelling & Insights product, and created a solution architecture for the System of Insight (SOI). In addition, Icon developed architecture patterns for building data pipelines to the SOI and established
an ECPP data reference architecture.
The target architectures were designed based on the principles driving the bank’s data technology strategy, for example, adopting data fabric and data mesh, ensuring a scalable and resilient infrastructure.
There were many beneficial features of the architectures, including federated ECP
Systems of Record (SoR), and an ECP System of Insight (SOI), storing ECP-domain data only. To enable near real-time insights, Icon defined patterns for creating pipelines between the SoR and SOI, using streaming where feasible, providing near real-time access to analytical data. The architectures also introduced cloud-based ML models for
management and deployment capabilities.
Icon established multiple patterns for the delivery of data products, both analytical and operational, to ensure that data scientists had access to the right data at the right time. In addition, patterns for accessing domain-SOIs, for example, payments or customer, were deployed to enable ML model development and training.
The solutions implemented by Icon were based on the following technology stack:
- Data Fusion – data pipelines
- Kafka – data streaming
- BigQuery – analytical data platform
- Vertex AI – model management & deployment
- Model Garden – model access
- DBT – data engineer managed pipelines
- Data Catalog – metadata management
This engagement shows how Icon’s expertise can help organisations tackle difficult challenges and deliver transformative outcomes. By defining a future-ready architecture for the bank’s Economic Crime Prevention Platform, including ensuring that the right data foundations were in place to fully leverage AI, Icon provided a blueprint that empowered the bank to enhance their fraud detection capabilities and maintain leadership in the economic crime prevention domain.”
Our unique contribution
Previously, there was insufficient prominence given to the bank’s data and technology strategy within ECPP, which Icon has helped the bank overcome to discover a clear technology strategy. A particular element of this was aligning both the bank technology and ECP strategies to deliver a coherent and consistent target architecture.
Icon supported the bank to define how target architectures should be modelled at both platform and product level. The bank utilised Icon’s expertise in AI & ML, data platforms and the cloud. For instance, Icon demonstrated to the bank how to harness available cloud services to create the optimal, most cost-effective AI, ML and data solutions whilst meeting ECP’s needs.
