The Operations AI Maturity Model
Five stages from "dashboarded" to "agentic." A self-assessment for operations leaders considering an AI-native platform.
Operations leaders ask us a version of the same question: "are we ready for agentic AI?" The honest answer is "depends on where you are." This paper proposes a maturity model — five stages — to make that question precise.
The five stages
Stage 1 — Dashboarded. Data is visible. Decisions are manual. The organization can see operations but can't intervene faster than the next standup.
Stage 2 — Alerted. Anomaly detection fires. Operators get pinged. The pace of intervention shortens from days to hours.
Stage 3 — Recommended. Models propose actions. Operators approve. Decision time compresses to minutes. Cognitive load increases.
Stage 4 — Supervised agentic. Agents act under policy with operator approval at thresholds. Routine decisions automate. Operator span of control grows.
Stage 5 — Autonomous agentic. Agents act independently within policy. Operators handle exceptions. The operating model is fundamentally different.
Self-assessment
The paper includes a 30-item self-assessment with scoring guidance. Most operations organizations sit between stages 2 and 3. The interesting question is what it takes to move.
The PDF is in preparation
The full paper is on this page now. Leave a work email and we will send the PDF edition the day it is published — nothing else.
Thank you — we will send it when it is ready.