Implementing Architecture-as-code to enable data-driven architecture

The client is the Commercial & Institutional (C&I) division of a UK Tier 1 bank, which serves all segments of UK business, from SMEs up to large corporate customers. The Chief Architect had set out to fundamentally disrupt and modernise the bank’s architecture methodology, with a vision to embed architecture into the end-to-end engineering lifecycle and apply the same scientific rigour as software engineering.

The bank’s existing approach to architecture was entirely manual: slow, and heavily dependent on static diagrams, making it subjective, inconsistent and open to interpretation.
To realise the target vision, the bank needed to move away from analogue processes and tools, and towards automated, data-centric practices – essentially turning traditional architecture artefacts into Architecture-as-Code.

Icon worked closely with the lead architect of each C&I domain to build a detailed understanding of existing  domain architectures and drivers. The team validated the quality and accuracy of existing architecture data, identifying and quantifying any gaps and ensuring the specific concerns, needs, and pain points of each domain were fully understood and addressed.

Building on this foundation, Icon aligned the Architecture-as-Code approach with the bank-wide architecture strategy process. A formal data schema was designed and then used to establish a clear picture of the current state and define a roadmap for future change, including each domain’s business outcomes and risks, the architecture outcomes required to support and satisfy them, and the business applications owned by each domain that would be impacted by those changes.

Icon then populated the code repository with architecture data baselines for each C&I domain, modelling each area to the level required to support strategic decision making. In some cases, this involved formally specifying architecture data for the first time. The data schema was also further extended to cover all aspects of architecture data, and then used to demonstrate modelled examples covering business capabilities and processes, application structures and integrations, and technical deployments mapped to underlying technologies and capabilities.

The client has fully embraced the Architecture-as-Code approach, using Icon’s data model to develop a formal data specification for Architecture-as-Code using their chosen technology, which brings together specification, modelling, and visualisation capabilities, and used this as the foundation for the architecture data repository. Based on this, the client has completed models for all technical domain architectures, linking and consuming data programmatically from existing architecture sources wherever possible.

In parallel, the client has implemented source code management practices for its codified architecture, enabling robust version control and governance. This has significantly streamlined the strategic architecture roadmap process for domain lead architects, reducing complexity, improving efficiency and accelerating decision-making. It has also introduced greater flexibility, allowing domain architects to use the tools that best suit their needs without compromising consistency.

The client has since showcased the benefits of this approach to other stakeholders across the organisation, positioning themselves as a leader in modern, data-centric, automated, architecture practices.

The client required a partner to accelerate its architecture transformation initiative. While there was a strong appetite for change within the architecture community, internal bandwidth was limited, and a structured approach was needed to ensure consistent alignment both centrally and across all domains.

Icon provided the consultancy required to lead the transformation across the division, along with the technical capability to implement an Architecture-as-Code repository fully aligned with the bank’s engineering practices.