Using Architecture-as-code to transform digital banking

The client is one of the UK’s largest retail banks, serving over 30 million customers and operating primarily in the UK.

One of the client’s largest and most business-critical technology platforms is its Digital platform, requiring six technical teams, and covering 180 customer journeys. This platform is currently undergoing a major modernisation programme, moving to cloud-based services.

The organisation was facing the challenge of a significantly increased workload which, due to the architectural complexity of the environment and ongoing large-scale legacy transformation, was expected to overwhelm current architecture practices. There was therefore a critical need for greater architecture efficiency and effectiveness to enable the platform to deliver its objectives.

Icon began by investigating and analysing the architecture methods and tools available across the organisation. This also included developing some manual Architecture-as-Code assets in parallel, to rapidly determine the most effective implementation approach. Following this, Icon chose to utilise Ardoq, the client’s incumbent strategic Enterprise Architecture tooling, as the most appropriate platform for the Architecture-as-Code deliverables, which were to be implemented as reports and dashboards about the Digital platform.

Icon established a codified data foundation built upon the Ardoq tooling, making use of Icon’s Architecture-as-Code Reference Model. This enabled the client to both expand the scope of the codified architecture and build on the baseline architecture views, as data quality was improved and more processes were embedded.

Icon also created an Architecture Baseline for all technical teams within the Digital domain, which has been codified as data in a single data store, hosted by Ardoq, and initially populated with business capability data. This data is linked to application landscape data across both organisational boundaries and data repositories. This was followed by the implementation of a Target State Reference Model, and a process for Architecture-as-Code for use both within the Digital platform and more widely across the bank.

Finally, Icon delivered pragmatic improvements to architecture processes to embed the use of Architecture-as-Code, including the use of patterns and design documentation, as well as making recommendations for further capability enhancements.

Icon delivered various architecture benefits, including controlling data entry in architecture processes, which has resulted in improved Digital architecture data quality. Icon also enabled reuse of automation in other domains across the bank through the cloning of Ardoq views and dashboards.

In addition to this, Icon correctly codified and linked architecture data for the first time in the Digital domain, increasing the Digital architecture delivery rate and reducing error risk. Icon also made it significantly easier to maintain the Architecture Baseline for Digital with process improvements.

The coverage of Architecture-as-Code will increase as data availability grows, providing a current and accurate view of the technology landscape that will increase confidence and usability for all consumers. Icon has implemented small process changes that have been integrated into existing workflows to ensure that architecture data is consistently kept up to date.

The team also established a consistent format to articulate the target state, which can be aligned with Enterprise Architecture roadmaps. The Digital reference architecture approach is fully specified, with revised architecture processes and codified Digital design templates and patterns to ensure consistent designs and architecture decisions.

Icon's pragmatism and expertise in architecture methodologies and Architecture-as-Code were both key to helping the client improve their Digital architecture operations. Immediate delivery of value was required, not just slides or a roadmap, and so making meaningful changes quickly was critical: Icon kick-started the initial architecture data population manually, and in parallel made local architecture process changes to accelerate adoption, thereby increasing both architecture data quality and coverage. The approach taken also considered the wider bank view, looking ahead towards long-term architecture practice improvements across the organisation.

Icon's client-first, vendor-neutral, approach was also important: rapid analysis to understand the architecture landscape and the changes needed, meant that the affirmation of the incumbent EA tooling could be justified objectively. Ardoq was confirmed as the right choice for the client's needs, primarily due to its reporting and dashboarding capabilities, and
also speed to deliver value.