the challenge.

The organization’s growth and modernization efforts were blocked by a combination of legacy technical debt, rigid operational constraints and limited access to critical domain expertise.

  • the inaccessibility of legacy calculation logic

    The previous closed, decades-old system whose business rules lived entirely in code, old documentation and tribal knowledge rather than in maintainable, documented services. Even small changes to coverage or premium logic require disproportionate effort. 

  • fragmented data and auditability gaps

    Reference data, such as premium rates, were scattered across the legacy landscape with limited documentation and inconsistent lineage. This created significant complexity for validating calculations and preparing historical data for migration required for the new platform.

  • the timeline challenge

    The client’s internal experts were managing multiple priorities, which limited their availability to address domain-specific challenges and provide input on discovery and requirements-gathering approaches within the project timeline.

  • architectural ambiguity and hotspots

    Certain structural dependencies within the legacy system threatened to derail the project. Without early alignment and validation, these dependencies created significant delivery and planning risk. 

the solution.

Randstad Digital delivered a solution that combined AI-assisted discovery workflows with architectural oversight and agile delivery governance.

  • modern azure-native architecture and MVP delivery

    We architected a greenfield stack on Microsoft Azure using .NET 10 with EF Core. During the MVP, the team identified that the initially planned services were too interdependent to work as separate standalone components and made the decision to pivot to a modular monolith, i.e., a real-time architectural call backed by documented trade-off analysis, and a hallmark of a mature delivery team. 

    We delivered a five-week MVP that modernized a representative slice of the calculation engine, proving the end-to-end stack, including serverless Azure SQL deployment, CI/CD pipeline automation, foundational calculation services and Salesforce integration, before the full-scale build.

  • strategic sprint 0 and hotspot resolution

    To prevent mid-project rework, we conducted a dedicated Sprint 0 to resolve all critical architectural dependencies before the full build began. Through joint event-storming sessions with the business and IT teams, the domain was decomposed into 10 distinct calculation services, establishing a validated build sequence, migration planning for historical normals, service ownership boundaries and yield experience transition strategies.

  • agile delivery and multi-phase build management

    The engagement transitioned from a five-week MVP phase into a full build phase running for three consecutive quarters, operating on three-week delivery cycles. A formal change control process ensures scope and budget remain tightly governed throughout delivery. The Delivery Lead who ran the MVP alone now leads the full team, meaning the domain knowledge, architectural decisions and client relationships built during the MVP transfer directly into the build phase with no ramp-up cost.

  • product owner management and quality governance

    Throughout the iterative build phase (spanning yield, coverage, premium and experience transfer), the team operates on product owner-led refinement and demos. Quality is managed through automated testing, regular code reviews and performance testing aligned to SLAs. To address the legacy documentation gap, all decisions are captured in a centralized Azure DevOps wiki, ensuring long-term maintainability after the final closeout.

  • built to enterprise security standards

    The platform was built with a security-first posture from day one. No credentials are stored anywhere in the system. Authentication runs through managed identities and Azure Key Vault. The entire platform sits inside a private virtual network with a deny-by-default inbound policy. Every request is automatically logged and traceable. Performance thresholds are enforced in the automated pipeline, not aspirational targets. And new calculation modules can be scaffolded in minutes from a built-in template, meaning the platform grows without being rewritten.

the results.

Key outcomes included:

  • accelerated architectural confidence: Successfully moved from an initial scoping discussion to fully functional working software on Azure in just five weeks, validating the complete .NET 10 / EF Core / Azure SQL stack end-to-end.
  • delivered under budget: The 5-week MVP phase came in approximately 40% under budget, with AI tooling eliminating the need for a full team during the discovery and scoping phase.
  • digitized logic and traceability: Replaced tribal knowledge with a documented backlog of ~323 story points across 93 items and 15 features, with all formulas and acceptance criteria traced directly to the official insurance manuals.
  • unified domain alignment: Achieved full consensus on the technical roadmap through joint event-storming sessions, resulting in a shared model of 10 distinct calculation services and a validated build sequence.
  • production-ready from day one: Zero stored credentials, automated CI/CD, multi-level testing including load and performance profiles, and full request-level observability were built in from the start, not added later.

ready to de-risk your digital transformation?

Explore how Randstad Digital helps enterprises transform their most challenging legacy environments into scalable, high-velocity operations.

contact us today