// System
Four components. One system.
The role only works when the surrounding organization gives it direction, leverage, escalation, and a place to hand work off.
// Interaction Diagram
How the system holds the role together.
The AC sits in the center, but the role is only viable because these surrounding components define how autonomy is directed and contained.
// Directs
Company Context
Defines what good judgment looks like and keeps autonomy aligned with strategy instead of personality.
// Expands Execution
AI Agents
Generates analysis, design, code, and operational output under AC direction while leaving authority with the operator.
Core Role
Autonomous Contributor
The AC is the decision-maker. The surrounding system gives them direction, execution support, escalation, and a place to hand work off.
// Resolves Escalations
Steering Board
Steps in when decisions exceed the AC mandate, collide with strategy, or require portfolio-level tradeoffs.
// Constrains and Receives Handoff
Platform Team
Owns architecture, shared systems, and long-term maintenance so shipped work does not fragment the product.
// Components
What each component contributes.
// Specialized operators
AI Agents
AI agents expand what one person can do. They execute work, surface options, and prepare material, but the AC remains the decision-maker and quality gate.
- — Execute work, propose approaches, and prepare decision material
- — Operate in three modes: Tool, Advisor, and Decision Support
- — Improve through structured feedback from retrospectives
- — Escalate to humans where verification requires domain expertise
// Shared platform and governance
Platform Team
Provides the infrastructure, architectural contract, and maintenance path that let ACs ship without fragmenting the product.
- — Complementary maintainer role that operates and sustains shipped outputs
- — APIs, monitoring, observability, databases, hardware resources
- — Data model ownership and management
- — Technical debt identification and prioritization
- — Product architecture decisions spanning AC projects
- — Platform evolution through better APIs, tooling, and automation
// Escalation and strategic control
Steering Board
Defines autonomy boundaries, resolves exceptions, evaluates performance, and manages portfolio-level choices. It is a strategic authority, not a routine approval layer.
- — Sets framework for AC autonomy through context, budget, and thresholds
- — Decides in escalated cases flagged by AI validation
- — Evaluates AC performance
- — Manages the project registry and portfolio priorities
- — Decides on sunsetting or continuing features
// Direction and constraints
Company Context
Vision, mission, personas, do's and don'ts, and strategic priorities. This context gives AI validation and human judgment the same frame of reference.
- — Defines direction and strategic priorities
- — Constrains AC autonomy within clear boundaries
- — Continuously refined from decision log and retrospectives
- — Guides AI validation of new ideas and exceptions
// Coordination
Coordination without adding a coordinator.
The model replaces some human coordination overhead with explicit system mechanisms. That is what lets autonomy scale past one unusually strong individual.
AI Detection of Overlaps
AI has access to the project registry, company context, and platform contract. When validating a new idea, it automatically checks for overlap with ongoing projects, conflict with planned projects, and impact on shared components.
Platform Team as Consistency Guardian
The platform team owns data models, customer journey, design system, and shared APIs. The AC doesn't need permission but must comply with architectural requirements that ensure their output fits the whole.
Context Sharing Among ACs
The form depends on scale. 1–3 ACs: AI + project registry is sufficient. 4–7: Regular async updates. 8+: Short weekly AC standup facilitated by AI.
// Lifecycle
Creation meets maintenance.
The AC creates new value. The platform team keeps that value reliable after the handoff. The system breaks if either side is missing.
Launch
AC delivers, platform takes over. Feature meets the contract, passed review gate. KPI measurement begins.
Operations
Platform maintains the feature in production. Regularly updates operational costs. Metrics compared against expected impact.
Review
Periodic evaluation: benefits vs. costs, user utilization, maintenance costs, strategic relevance.
Decision
Steering Board decides: continue (delivering value), iterate (needs modifications), or sunset (insufficient value).
Platform capacity is an explicit system constraint. If nothing can be handed off safely, nothing new should be accepted casually.
// Sunset
Sunsetting is capacity management.
A feature leaving the system is not a failure ritual. It is the mechanism that keeps the platform from becoming a graveyard of low-value maintenance work.
- — Steering Board decides on sunset based on data
- — Platform team plans shutdown timeline (data migration, user notifications, dependency removal)
- — AC who created the feature may be informed but has no veto
- — Released capacity returns to the system and enables acceptance of a new project