AI Company Framework
How an AI-operated organisation is structured, department by department: responsibilities, KPIs, agents, SOPs, templates and workflows.
Most companies keep their operating model in a document nobody opens. We publish ours. Each department page states what it owns, the numbers it is judged on, which work is handled by an AI agent and which deliberately is not, and the workflows that connect it to everything else. It is written to be used, and to be copied.
CEO
Direction, capital allocation and the decisions that cannot be delegated.
COO
Operating cadence, execution and cross-function throughput.
CTO
Technology strategy, architecture authority and build-vs-buy.
PMO
Portfolio, delivery governance and dependency management.
HR
Hiring, capability, performance and the human side of automation.
Finance
Planning, unit economics and cost attribution.
Sales
Pipeline, qualification and forecast discipline.
Marketing
Positioning, demand generation and attribution.
Legal
Contracts, IP and AI-specific liability.
Compliance
Obligations, evidence and audit readiness.
Security
Posture, controls and incident response.
QA
Quality strategy across products and agents.
DevOps
Delivery pipeline and operational reliability.
Engineering
Build capacity, standards and technical debt.
Support
Service levels, escalation and feedback into product.
AI Strategy & Transformation
Where AI is applied, where it is refused, and whether the change actually paid.
AI Engineering & Product
Evaluation, guardrails, cost per outcome and how much a feature may do unsupervised.
Data & Analytics
One authoritative number per concept, its lineage, and the gates before it is published.
Research & Innovation
Cheap experiments against written kill criteria, and the negative results worth keeping.
Knowledge Management
One authority per topic, active retirement, and the context every agent reads.