Frameworks

Execution Integrity is the degree to which an organisation's execution system maintains predictable, verifiable output quality as AI involvement increases. It is not a cultural attribute — it is a measurable structural property. An organisation with high Execution Integrity can demonstrate that its outputs are human-verified, its drift is contained, and its accountability structures are functioning. Low Execution Integrity manifests as silent failures, unverified AI outputs, and accountability gaps that compound over time.

CGEI is the formal research framework that explains why AI-assisted execution degrades without structural human governance. The CGEI anchor theorem states: execution integrity is a function of human verification density, correction latency, and accountability structure fidelity. CGEI distinguishes between execution environments with durable structural integrity (GLASS substrate) and those that collapse under operational stress (SAND substrate). Co-Regulated Execution Systems are the operational delivery model derived from CGEI.

Co-Regulated Execution Systems describe the governance and delivery model in which humans and AI actively co-regulate execution quality in real time. Rather than AI systems operating autonomously with periodic human oversight, Co-Regulated Execution requires continuous, structured human verification at defined checkpoints. The operational delivery unit is the Execution Pod — a 3–4 person human team + AI that maintains Execution Integrity across a bounded scope of work.

Human Machine Intelligence is the overarching model for human-AI collaboration that goes beyond tool use to describe the structural relationship between human execution capability and AI-augmented output. HMI identifies the conditions under which human-AI teams outperform either alone, and the failure modes that emerge when AI involvement erodes the human verification layer. Execution Pods are the delivery unit derived from HMI.

GLASS and SAND are substrate classifications for AI-assisted execution environments, derived from the CGEI framework. A GLASS substrate is structurally stable: human verification is embedded, accountability structures are clear, and correction latency is low. A SAND substrate appears functional but lacks structural integrity — it degrades under operational stress, and failures cascade silently. The GLASS/SAND distinction is the technical framing for Execution Survivability. Classification methodology lives at Developer Insights (PNT).

The EI Score is the measurement instrument for Execution Integrity. It scores organisations across five structural dimensions: Human Verification Density (D1), Inspectability (D2), Accountability Structure Fidelity (D3), Execution Drift Control (D4), and Correction Latency (D5). Each dimension is scored 1–10 and combined into a composite Tier A–D rating. The specification is permanently hosted at aiadoptionperformance.com; the live diagnostic implementation runs at the same domain. Co-authored by Duena Blomstrom and Dave Ballantyne.