Enterprise applications have traditionally provided a useful boundary for understanding architecture. A system often supported a recognizable business area, held significant parts of its data and had a relatively clear operational owner.
That boundary is becoming less useful as the landscape becomes more composable.
A customer journey may cross a SaaS platform, proprietary services, identity providers, data products and several APIs. An AI-enabled capability may depend on information governed elsewhere and on controls that do not appear in its user interface. The application remains visible. The conditions that make it work are distributed.
When a capability is assembled across several services, the architecture question changes. It is no longer only “what does this system do?” It becomes “what must remain true between these systems for the capability to operate?”
The systems still matter. But increasingly, the risk and consequence of change sit in the relationships between them.
That includes data meaning, interface behaviour, identity, ownership, service levels, regulatory evidence and the sequence in which a change can be introduced. Much of this knowledge is represented somewhere, but rarely in one place.
The architecture repository may describe applications and capabilities. The CMDB may describe deployed components. An API catalogue may describe interfaces. Documents explain decisions and exceptions. Experienced people know which relationships are fragile. Each source sees part of the enterprise.
The most important architecture knowledge is therefore increasingly relational. It lies between the records.
Composable platforms are attractive because they increase choice. Organizations can retain stable products, combine services and build selectively where differentiation matters. But choice transfers work from the individual product to the architecture around it.
The organization must know how the selected parts fit, who is responsible for the joins and what evidence supports the design. Otherwise composition becomes accidental complexity hidden behind a portfolio diagram.
This is why architectural context grows in importance as application portfolios become more flexible. The need for architecture does not disappear when platforms become easier to connect. It moves toward the connections.
The answer is not necessarily to copy every source into a larger repository. That can create a new concentration of data while losing provenance and ownership.
A better starting point is to make relevant knowledge accessible in context and connect it around a real change question. Preserve the source. Retain the route back to it. Make the relationships inspectable. Apply the access rules of the environment in which the information belongs.
This is the practical territory of ARC. It begins with the architecture environment and knowledge the organization already has. Access releases useful material from particular tools and formats. Connect brings selected sources into a common decision context. Insight helps architects inspect relationships, gaps and consequences.
The objective is not a perfect digital copy of the enterprise. It is a sufficiently trustworthy operating view for the next controlled change.
As landscapes become more composable, architecture leaders should worry less about creating one universal source of truth and focus more on knowing which sources are authoritative for what — and how their truths can be related responsibly.
The future application portfolio may be assembled from many platforms. Its execution capability will depend on what the organization understands between them.