Building for the Enterprise
Tie everything together — architecture reviews, multi-team coordination, CI/CD strategies, and delivering ongoing value.
1 · Multi-Team Architecture
At enterprise scale, multiple teams build on the same Foundry instance. A good architecture separates concerns:
- Shared layers — central datasets and Object Types that everyone depends on (e.g., Employee, Product, Location).
- Domain layers — team-specific pipelines and applications (e.g., Supply Chain, Finance, HR).
- Consumption layers — published Workshop apps and API endpoints.
Use project boundaries and markings to enforce isolation while enabling collaboration.
2 · CI/CD Strategies
For production-grade deployments, establish:
- Branch protection rules — require CI checks and peer review before merge.
- Staging environments — test changes on realistic data before promoting to production.
- Automated regression testing — build comparison reports that flag unexpected changes in output data.
- Release management — tag releases and maintain changelogs so rollbacks are straightforward.
3 · Delivering Ongoing Value
The final challenge is not building — it's sustaining. Enterprises succeed with Foundry when they:
- Measure adoption — track how many users interact with Ontology apps weekly.
- Run office hours — regular sessions where the platform team helps domain teams.
- Maintain a backlog — prioritise Ontology improvements based on user feedback.
- Celebrate wins — share case studies of time saved and decisions improved.
Foundry is a platform, not a project. Treat it as a living system that evolves with the business.
✅ Key Takeaways
- Separate shared, domain, and consumption layers for multi-team success.
- CI/CD with branch protection, staging, and regression testing ensures quality.
- Measure adoption, run office hours, and maintain a backlog for continuous value.
- Foundry is a living platform — invest in ongoing evolution, not just initial build.