// Framework
Reference appendix.
Decision rights, decision logging, metrics, and accountability structures for teams already evaluating or adopting the model.
// How to use this page
This page is intentionally reference-heavy. Start with the model and adoption pages if you are new to Autonomous Contributor, then use this appendix when you need the operating details.
// Decision Rights
Three decision zones
Most AC decisions go through AI validation — not through the Steering Board. The better the company context and decision log, the fewer escalations arise.
AC Decides Alone
The decision aligns with company context, fits the budget, respects the platform contract, and does not interfere with other projects.
The AC proceeds immediately.
- — Launching a project within budget
- — Minor changes to an existing feature
- — Technology selection within the contract
- — Iteration based on data
AI Validates with Warning
The decision approaches a threshold — budget limit, partial overlap, or moderate user impact.
AI approves but records a warning visible to the Steering Board.
- — Project near the budget ceiling
- — Change affecting a specific user segment
- — Use of technology outside the standard stack
Steering Board Decides
The decision affects strategy, exceeds budget limits, impacts a large user base, or conflicts with another initiative.
Escalation is mandatory.
- — New product or business line
- — Pricing model change
- — Decision with legal or regulatory impact
- — Project requiring resources beyond AC budget
| ACs | Without Decision Rights | With Decision Rights |
|---|---|---|
| 1–3 | SB can handle everything | Works, but not necessary |
| 5–10 | SB becomes a bottleneck | Most decisions in green zone, handful of escalations monthly |
| 10+ | Model collapses | AI validation scales linearly; SB load grows minimally |
// Decision Log
The organization's memory
The central decision log serves as the organization's memory. Without it, the system cannot learn and AI cannot improve its recommendations.
At Creation
Filled by AC
- — Decision ID — unique identifier
- — AC — who made the decision
- — Date — when
- — Decision type — new project / change / sunset / pivot / technology choice
- — Problem or opportunity — what was identified and why it's important
- — Chosen solution — what was done and why this approach
- — Rejected alternatives — other approaches considered and why discarded
- — AI recommendation — what AI recommended and whether AC followed it
- — Assumptions — what the decision rests on
- — Expected impact — metrics that should improve, and by how much
- — Budget — how much the decision costs
- — Decision zone — green / yellow / red
- — Dependencies — which other projects this affects
After Delivery
AC Retrospective
- — Actual impact — what actually happened with the metrics
- — What worked — which parts delivered value
- — What didn't work — which parts failed and why
- — AI agent quality — where AI helped, where it failed
- — Lessons learned — what would be done differently
Long-Term
AI + Steering Board
- — Short-term result (1–3 months) — is it working?
- — Long-term result (6–12 months) — still working? Side effects?
- — Assumption validation — which assumptions were confirmed or not
- — Attribution — what drove the result
Principle: If filling in a record takes longer than 15 minutes, the structure is too complex. The AI agent should pre-fill 80% of the fields.
// Success Metrics
What gets measured
A two-tier system: system metrics (same for all ACs) and project metrics (defined per project).
System Metrics
Delivery Rate
How many projects the AC shipped to production per period.
Hit Rate
Share of delivered projects that achieved expected impact.
Time to Value
Average time from problem identification to measurable impact.
Estimate Accuracy
How close the ROI estimate at validation was to the actual result.
Handoff Quality
How many projects the platform team accepted without rejection.
Cost vs. Benefit
Aggregated ratio of AC budget + operational costs vs. benefits.
Project Metrics
Revenue / Growth
- — Revenue increase
- — Conversion rate
- — Number of new customers/users
- — ARPU
Retention / Engagement
- — Retention (daily, weekly, monthly)
- — Churn rate
- — Engagement metrics
- — NPS or CSAT
Efficiency / Cost Saving
- — Time saved (hours or FTE)
- — Cost reduction
- — Automation rate
Strategic / Infrastructure
- — Adoption rate
- — Internal user satisfaction
- — Impact on future project speed
- — Reputational benefit
Rule: No project without a metric. A project without a defined success metric at validation cannot proceed to execution.
// Accountability
Autonomy within a framework
Company Context
Vision, mission, personas, and do's & don'ts define the space in which the AC operates. Continuously refined based on the decision log.
Validation Gate
Mandatory research and idea validation. AI validation determines whether the AC proceeds alone or escalates.
Budget Control
The AC has their own project budget. Projects within budget are in the green zone — launched independently.
Review Gates
Each company sets its own review gates. This doesn't limit AC autonomy — it defines the conditions for it.
Platform Contract
Clearly defines technical and procedural requirements. Valid across projects. The AC has the right to negotiate exceptions.
Performance Evaluation
Evaluated by the Steering Board. Failure is data for iteration — repeated failure is a signal for change.
// Organizational Politics
The AC coexists with management
The AC doesn't replace management. Management manages people, processes, budgets, and long-term strategy. The AC delivers specific results end-to-end within their mandate.
The best solution to politics is prevention: transparency (all projects visible), clear mandate (what the AC can and cannot do), Steering Board as arbiter (decides based on data), and culture (the model works only when the company wants it to work).
// Strategic Coherence
Coherent, not reactive
Company context must contain current strategic priorities — specific areas where the company wants to grow and areas it doesn't want to invest in. AI validation checks every idea against these priorities.
The Steering Board quarterly reviews the portfolio and evaluates: Do AC projects cover strategic priorities? Are there uncovered areas? Is the portfolio balanced (quick wins, medium-term, strategic bets)?