12. Maturity Model

This section is informative. It describes maturity levels and states no requirement: no SHALL appears in it, and no conformance claim is assessed against it. Three things would have to be settled before it could carry one, and none of them is settled by measurement this Standard holds: the model does not describe Foundation at all (12.1, Level 2 says so outright); it gives no rule for aggregating dimension levels into an organizational level, so an organization at L3 on four dimensions and L1 on three has no stated level; and the dimension table stops at L3 while the levels run to L5. Stating a level therefore remains an organization’s own reading and is not a claim under 13.1.

12.1 Levels

Level 1: Ad Hoc

AI used opportunistically. No tasks, no gates, no metrics, no documented knowledge.

Level 2: Supervised (aligns with SENAR Core)

Tasks before work. AI output verified. QG-0 and QG-2 enforced. Core’s 2 metrics tracked — FPSR and Dead End Rate, which are the two SENAR Core names in its own Metrics section. Dead Ends documented (>15 min threshold). Session management is not part of this level: Core assigns the structured start and end of a work session to Foundation, and this model does not yet describe Foundation.

Key indicator: Every AI implementation has a Task; output is verified.

Level 3: Measured (aligns with Team)

All 5 responsibility sets covered. All 5 gates automated. All 10 metrics with baselines and targets (including ADR tracked per task). Knowledge capture structured. Federation active. Architectural Decision Records tracked — the abbreviation ADR is not used for them in this Standard, where 3.36 gives it to Adversarial Detection Rate. Data-driven decisions.

Key indicator: Baselines established; metrics reviewed at every Retrospective.

Note: Levels 4 and 5 are aspirational targets based on industry patterns (CMMI, DORA). No SENAR implementation has been independently validated at these levels. They are provided as directional guidance for organizations seeking continuous improvement.

Level 4: Managed (aspirational)

Cost predictability consistent. FPSR improving. Defect Escape Rate below threshold. Knowledge base actively improves AI context quality. Process improvements are experimental and measured.

Level 5: Optimizing (aspirational)

Organization defines own patterns. Value streams optimized. Process experiments routine. Knowledge base is competitive advantage.

NOTE: Levels 4–5 are aspirational. See the note above Level 4 for details.

12.2 Assessment Dimensions

Organizations MAY be at different levels across dimensions:

12.3 Progression

Organizations SHOULD progress sequentially. Each level builds on the previous. Organizations MAY focus on weakest dimensions first.