In a volatile sourcing environment, procurement intelligence helps organizations make sense of supplier risk, market shifts and cost pressures. For Beroe, a company built around delivering that intelligence to enterprise customers, the strength of its platform was central to the value it provided.
Over time, that platform expanded to support more capabilities, workflows and customer needs, adding complexity to the technology environment behind it. The company wanted to simplify its technology foundation and make it easier to maintain, scale and evolve, while ensuring customers could continue using the platform without disruption.
Built for Growth. Ready for Simplification.
The company’s existing platform was already delivering value. The existing workflows, integrations and user journeys supported customers’ day-to-day needs.
However, over time, the technology environment had become more complex. A focused platform had expanded into a broader ecosystem of applications, services and data connections, supported by Java-based backend services and a front-end split across Angular and React. Maintaining multiple technology tracks meant different skill sets and increased development effort, making future enhancements more difficult.
Beroe saw an opportunity to simplify the stack by reducing operational complexity, improving maintainability and creating a cleaner foundation for future innovation, including the addition of AI-driven capabilities.
Modernizing a Live Platform
The goal was a like-for-like modernization. Backend services would move from Java to Python and remaining Angular components would transition to React. Existing functionality, API contracts, integrations and user journeys would all remain intact.
The work involved a live platform that continued to support customer workflows while modernization progressed. This meant that the company needed to have a detailed understanding of the platform’s business logic, dependencies and architecture. Years of product development had created a rich set of workflows and rules that had to be preserved through the migration.
Persistent worked with Beroe to plan the modernization so as to reduce risk and maintain continuity. Each release had to support the current platform experience and move the technology foundation toward a simpler, more scalable architecture.
A More Structured Approach to Modernization
Persistent worked with Beroe to bring structure and speed to modernization. The team began by analyzing the existing platform, mapping dependencies and identifying areas where migration could be accelerated without increasing the delivery risk. This helped create a clearer view of the work ahead.
Further, instead of treating modernization as a direct technology conversion, the team focused on improving maintainability, addressing technical debt and strengthening quality across the platform.
The company had already explored using AI in parts of its migration effort. Persistent extended that early experimentation into a more structured delivery model, with clear checkpoints for analysis, migration, review and validation.
Services were assessed independently and moved through defined delivery lanes. Automated runs progressed work overnight, while morning reviews helped the team validate outputs, identify blockers and decide where human judgment was required.
Using AI to Accelerate Engineering Delivery
To support the structured migration model, Persistent incorporated Anthropic’s Claude into the modernization effort using Beroe’s existing licenses. The goal was to help engineering teams with repository analysis, dependency mapping, migration planning, code migration and quality review.
Claude was particularly useful because the engagement required careful like-for-like migration rather than broad functional redesign. It helped reduce the manual effort involved in understanding and migrating complex application components, while keeping visibility into the business logic, dependencies and workflows that needed to be preserved.
This allowed architects and developers to spend more time on design decisions, validation and problem-solving. Human review remained central to the process, with architecture sign-offs, backlog readiness checks, implementation reviews and release approvals built into the delivery model.
The model also emphasized traceability. Automated workflows worked from verified inputs and systems of record, stopping when information was unclear rather than making assumptions.
For Beroe, this made AI a practical enabler of modernization: helping accelerate delivery while keeping quality, control and continuity at the center of the program.
Building for What’s Next
The modernization program is already delivering measurable progress. The first release has been completed successfully; additional releases are planned and four of the 19 repositories in scope have already been migrated.
Beyond the migration itself, the engagement is also improving the quality and maintainability of the platform. As teams worked through the codebase, they addressed technical debt, strengthened testing practices and improved code quality. Security findings identified during assessments were remediated and test coverage improved significantly across migrated components. In one example discussed during the engagement, test coverage increased from effectively zero to approximately 99%, exceeding the committed coverage threshold for migrated components.
On the frontend, the work is also helping improve performance, component reuse and module-level maintainability within the boundaries of the like-for-like migration. For the company, these improvements matter because success was never defined only by moving from Java to Python or from Angular to React. The larger goal was to simplify the technology landscape supporting the business, reduce complexity and make the platform easier to maintain, extend and evolve.
The modernization effort is still underway, with releases progressing in phases, but it is already creating a stronger foundation for future AI-driven initiatives and product innovation.
