open-ai-transformation-maturity-model

Home · Getting Started · Model · Assessment Hub · Roadmap · References


Pillar 1 — Intent & Specification Fidelity

License: CC BY-SA 4.0
https://creativecommons.org/licenses/by-sa/4.0/

This pillar assesses whether an organization can express desired system behavior, constraints, and expected outcomes with sufficient precision for AI systems to implement safely and reliably.

👉 Back to Assessment Framework
👉 Back to Unified Scoring System


Why This Pillar Matters

As organizations move toward higher levels of AI transformation maturity, the limiting factor shifts away from code generation and toward the quality of intent expression.

At higher maturity levels, specifications become the control plane for software delivery. If intent is ambiguous, incomplete, or inconsistent, agentic systems amplify those weaknesses at scale.


What This Pillar Evaluates

This pillar evaluates whether the organization can:


Typical Assessment Questions

Examples of criteria within this pillar include:


Evidence to Look For

Useful evidence may include:


Low-Maturity Pattern

At low maturity, organizations tend to rely on:

In such environments, AI can generate output, but not with consistently reliable alignment to intent.


High-Maturity Pattern

At high maturity, organizations tend to exhibit:

These characteristics are essential for safe progression toward spec-driven development and autonomous delivery.


Common Failure Modes

Ambiguous Intent

Specifications are present but too vague to support reliable implementation.

Specification Drift

Documentation, architecture, and actual system behavior diverge over time.

Documentation Theater

Organizations produce large volumes of documentation without making intent enforceable or machine-consumable.


Relationship to Other Pillars

This pillar strongly influences:


Scoring Interpretation

Score 0

Intent is mostly implicit, informal, or absent.

Score 1

Some specifications exist, but they are inconsistent, incomplete, or weakly enforced.

Score 2

Teams can define intent repeatably in some contexts, but quality varies significantly.

Score 3

Intent is structured, reviewed, and sufficiently precise across meaningful parts of the organization.

Score 4

Intent is explicit, machine-consumable where appropriate, consistently governed, and continuously improved.


Relevance by Level


↑ Back to top · Home