Context
A healthcare SaaS company was transitioning from startup-style custom client delivery toward a scalable product operating model following a private-equity acquisition.
The application estate had grown across 530 repositories, approximately 500 component repositories, 40 priority services, 83 APIs, approximately 2,000 endpoints, 11 customer-facing applications, and 15 engineering or delivery teams.
Challenge
System knowledge was distributed across code, documents, teams, Jira, Confluence, diagrams, and tribal knowledge. Leadership and engineering needed a consistent way to understand product boundaries, integration dependencies, modernization risk, and delivery impact.
My role
Lead the creation of a multi-level enterprise architecture model and connect it to modernization planning, product decisions, onboarding, client delivery, and engineering execution.
Approach
- Created estate-wide architecture views.
- Created per-product views.
- Mapped integrations and data movement.
- Created workflow-level views for critical processes.
- Connected higher-level architecture to code and repository context.
- Created views appropriate for executives, product teams, architects, engineers, and onboarding.
- Used architecture artifacts as inputs to modernization sequencing and risk assessment.
Outcome
- Established a shared model of the application estate.
- Improved role-focused onboarding from approximately two weeks to one day.
- Reduced architecture and high-level-design refresh cycles from annual manual work to approximately four hours.
- Supported modernization planning across the full estate.
- Improved visibility into product boundaries, integrations, dependencies, and client-delivery concerns.
- Influenced architecture and modernization practices across 15 teams.
The result was not merely better documentation. It was an architectural operating model that enabled more consistent modernization, delivery, onboarding, and product decisions.
What I learned
- Shared models create decision leverage only when they map to real delivery work.
- Audience-specific views matter as much as technical completeness.