Findings
Every layer runs automatic checks (deterministic validators). Any gaps or inconsistencies appear as
Second opinion
Beyond the automatic checks, the Second opinion asks an AI to critique your model for quality —
Fixing concerns
Every Second-opinion finding offers a way to fix it for real — so the concern is gone, not hidden. There are
Progress & loops
A re-review returns a fresh list of findings. On its own, "4 findings became 5" tells you nothing —
Ignoring concerns
Not every concern needs a fix. Some are deliberate choices, or things you can't address yet. Ignore
Protecting your fixes
Hand-made changes — a surgical Fix, an edited entity field, a moved capability — are stored as
Narrative sync
Hand-made fixes land in the model but not in the Business Narrative, so the prose can silently fall behind
Prompt & output
Every modelling stage — Capabilities, Business Areas, Entities, Behaviour, Automations, Roles, Workflows,