01 Outcome clarityHow clearly is the operating problem and intended outcome defined?
The requested solution is named, but the underlying problem is still uncertain. The problem and users are known, but outcomes or exclusions need work. The problem, users, outcomes and explicit exclusions are documented.
02 Workflow evidenceHow well are users, decisions, handoffs and exceptions understood?
Mostly assumptions or screen ideas. The main path is known; exceptions and ownership are incomplete. Normal, exception and approval paths have named owners and evidence.
03 Architecture boundariesAre system, data and integration boundaries reviewable?
Key systems and data authority are unclear. Major components are known; interfaces or failure paths need decisions. System ownership, data authority, interfaces and failure behavior are recorded.
04 Decision ownershipWho can approve scope, data, security, acceptance and release decisions?
Ownership changes by conversation. Some owners are known, but important decisions lack authority. Material decisions have named owners and escalation paths.
05 Delivery controlCan the team produce and review working increments safely?
Work is large-batch or progress is mainly reported verbally. Increments exist, but review evidence or environments are inconsistent. Small increments, source control, environments and review evidence are routine.
06 Acceptance evidenceIs success expressed as observable behavior rather than a feature list?
Acceptance is informal or feature-based. Happy-path behavior is covered; permissions, failures or reconciliation need work. Positive, negative, access, data and operational acceptance are explicit.
07 Release readinessHow controlled is deployment, recovery and handover?
Release and recovery depend on individual knowledge. A release path exists; rollback, monitoring or handover is incomplete. Release decision, recovery, documentation and operational ownership are reviewable.
08 Operating feedbackDoes production evidence shape the next delivery decision?
Issues and requests arrive without consistent evidence. Some signals exist, but prioritization and ownership are inconsistent. Health, behavior, support evidence and review triggers shape the backlog.