Coherence, not level

The instinct with any maturity model is to chase the number — to treat L5 as the goal and L2 as failure. This assessment is built on the opposite conviction: coherence between a team and its habitat matters more than the level either has reached.

A coherent L2 beats an incoherent L4

A team sitting at L2 cognitive / L2 operational is healthier than one at L4 cognitive / L1 operational. The first team's thinking and tooling are in step: its practices are supported by its environment, and it can move up together. The second team's thinking has outrun its habitat — its specs and ambitions rest on tooling that can't enforce them, so the gap shows up later as drift, rework, and "the AI keeps breaking things".

This is why the Habitat Build Gap reports a regime, not just a magnitude. The regime tells you which way to invest; the magnitude alone would just tempt you to chase levels.

The weakest discipline is the ceiling

The same conviction governs the cognitive scoring heuristic: the assessed level is the highest level where the team has substantial evidence across all three disciplines, and the weakest discipline is the ceiling.

Strong specs with weak verification is L2, not L4 — because specifications you can't verify are aspiration, not control. Strong guardrails with no encoded context is L2, not L3 — because guardrails without shared context catch symptoms, not causes. Refusing to average across disciplines is what stops the score from flattering a team into believing it's further along than it is.

What this means in practice

• Don't read a low level as failure. Read an incoherent profile as the warning.
• Lift the weakest dimension or discipline, not the headline number.
• Re-assess over time and watch the gap close — coherence is the thing you're optimising for.

The framework's other ideas — the amplifier principle, the human pace — all assume coherence as the baseline. The Habitat Build Gap is the early warning that coherence has slipped.