Test-driven development (specification-driven verification)

Most teams write tests. Fewer write them first, in the classic TDD sense — and almost none have solved the deeper problem this dimension actually measures: whether verification coverage can keep scaling as generation volume scales, or whether it quietly becomes the bottleneck the moment an AI agent starts producing code faster than tests can be written by hand.

Where most organizations start (Nascent)

Tests exist and are written close in time to the code they cover, but authored independently, after implementation — the test verifies what the code happens to do, rather than the code being written to satisfy a pre-specified test. The first real step is moving test authoring ahead of implementation for defined classes of work, even if the specification and the test are still separate artifacts requiring manual translation.

Where the real gains happen (Modeled → Integral)

The meaningful shift is deriving tests directly from the specification itself, through a consistent, repeatable method — closing the translation gap so the same well-formed specification reliably produces the same tests, not a manual reinterpretation each time. From there, the bigger change is generating tests concurrently with the specification, so ambiguities and missing edge cases surface while the requirement is still being written, not after implementation has already begun.

What the top of the curve actually looks like (Telemetric)

At full maturity, specification authoring, test generation, and prototyping fuse into a single governed motion. Verification coverage scales automatically as generation volume scales — the system doesn't care whether work begins from a specification, a prototype, or a test; whichever artifact starts it, the others follow, staying aligned with each other.

Why this dimension matters

AI-assisted generation can produce code far faster than a human reviewer can meaningfully verify by hand — this dimension is what keeps that speed from becoming a coverage gap instead of a genuine capability gain.

---

Drafted from the SDLC model's real locked content — including the transition and verification notes now folded into ai_native_sdlc_maturity_model.md itself (2026-07-27). D7 carries no open review flag.