Architecture
The Architecture tab is where the design itself lives. One button runs the automated design loop: the platform generates an architecture from your inputs and the approved technology catalog, analyzes it against your controls, remediates what it can, and pauses only when a human decision is genuinely needed. The result is a component-and-flow model, not a drawing: every diagram, compliance check, cost figure, and delivery plan downstream is derived from it.
Two ideas make this tab different from a design canvas:
- Decisions are recorded, not hoped for. When you approve a technology for an open gap, that decision is platform state; the next run selects it deterministically no matter what the model does. When you edit or lock a component, your override survives re-runs.
- The boundary is governed. Re-runs cannot silently grow the authorization boundary. Additions must cite the control, finding, or requirement that forces them, and uncited additions are quarantined for your explicit accept-or-discard decision.
The tab has two segments: Design (components, flows, trust zones, the run panel, and the scope guard) and Decisions & Value (ADRs, cost, estimated value delivered, evidence, and exports).
In This Section
| Page | What It Covers |
|---|---|
| Run the Automated Design | The full loop: generating, decisions needed, applying picks, re-running |
| Decide Open Gaps | Choosing technologies for capabilities the catalog does not cover yet |
| Keep the Scope Stable | The scope guard, preserve-scope runs, and accepting or discarding proposed additions |
| Edit and Lock the Design | Manual overrides that survive AI re-runs |
| Design Chat | Q&A, ADR capture, and apply-able changes (Medium plan and above) |
| Decisions and ADRs | The decision record the generator and chat maintain |
| Cost and Estimated Value | Indicative run cost and the ROI value model |
| Evidence and Exports | Evidence capture and document exports |
| Business Case | The funding-decision PDF assembled from the package |