Glossary

Every Rocket Routine OS term in one place — explained briefly.

Also see: Strategy & Leadership Wiki

A B C D F G H I K L M O P R S T V W

A

Actor Registry: Record of every actor (human, AI operator, Knowledge Persona) working in the system, with its category, Role Contract, and operating domains. "If an actor isn't in the Registry, it isn't working in the system."
Adoption Levels: Three-stage ladder (Shadow → Copilot → Autopilot) an AI operator climbs per role, per domain, based on measured FTT evidence rather than trust.
AI Operator: An Actor Registry entity type: provisioned via Role Contract, runs routines, produces standardized outputs; bound by explicit decision rights, tool access, and escalation triggers. Carries execution load, not decision accountability.
Autopilot: Highest Adoption Level: the AI operator executes and ships within its Role Contract's hard boundaries; humans lead by exception and audit.

B

Branch: Decision Impact Classification tier for operational decisions (how a single domain runs); owned by domain leads.
Build Log: The monthly bilingual email (never called "Newsletter" in copy); reports the month's article of the month, key takeaways, a Company 0 update, and what's next.

C

Cadence: The five nested rhythm layers (Daily, Weekly, Monthly, Quarterly, Annual) that verify a different horizon of the I2I loop; a structural requirement, not an optional meeting calendar.
CEO-Bench: Princeton benchmark (ceobench.com) that gave 13 AI models and one rule-based baseline $1M each to run a simulated company for 500 days, used as evidence that autonomy without governance fails unpredictably.
Company 0: Rocket Routine GmbH itself, run live on Rocket Routine OS as the first production environment and proof point (always written with a space, never "Company0").
Constitution: A short set of enforceable rules the system actively checks and blocks violations of, as opposed to written "values" that merely sit on a wall.
Constitution Lite: The starting, minimal form of the Constitution used from day one: a handful of core rules plus a map of who decides what.
Control Tower: The single source of truth for work in flight; an event-driven Kanban-style state machine with pull logic, dependency visibility, and artifact linkage. Capitalized always.
Copilot: Middle Adoption Level: the AI operator executes within approval gates; a human approves every handoff before it ships. Written as one word.

D

Decision Impact Classification: The four-tier tree (Root, Trunk, Branch, Leaf) that sorts every decision by impact and assigns it a fixed owner.
Decision Moderator: The mechanism that formally selects a decision's method and is the sole invoker of Knowledge Personas; records their input but leaves the final call with the accountable Human Actor.
Domain: One of eleven fixed, value-producing units that together map any company completely, independent of headcount; distinct from a "department" (an org-chart/people concept).
Domain Advisory Board: Multiple Knowledge Personas combined within one domain, giving the Decision Moderator a synthesized (not smoothed-over) view.

F

Founder Mode: Operating Mode in which the CEO decides, executes, reviews, and corrects everything personally; AI operators only assist.
FTT (First Time Through): The primary quality metric: percentage of work items passing quality confirmation on the first attempt without rework. Always written in all caps.

G

Growth Mode: Operating Mode in which AI operators run routines and the CEO leads by exception; decision rights are explicitly distributed.

H

Human Actor: An Actor Registry entity type: holds decision rights, accountability, and judgment; holds every decision right not explicitly delegated; the only sovereign actor under the Constitution.

I

I2I Loop (Intent to Impact): The universal four-phase control loop — Intent, Insight, Implementation, Impact — structurally aligned with PDCA but requiring the system itself to change on divergence.
Impact: The fourth I2I phase: verifying quality and writing Learning that updates the system when reality diverges from Intent.
Implementation: The third I2I phase: executing routines, routed through the Control Tower, linked back to the originating Intent.
Initiatives: OMPRIKL artifact type: time-bound efforts that change the status quo, as distinct from Routines that operate the system as-is.
Insight: The second I2I phase: explicitly selecting the decision method (data, rule, procedure) before work begins.
Intent: The first I2I phase: defining outcomes and constraints as verifiable intent with explicit boundaries.

K

Knowledge: OMPRIKL artifact type: documented, verified information the system treats as trustworthy; anything outdated is treated as noise.
Knowledge Persona: An Actor Registry entity type: advisory-only AI Operator with read-only access to one specific published knowledge source; no write access, no decision rights, no routine ownership; invocable only by the Decision Moderator. "A Knowledge Persona speaks. It does not decide."

L

Launchpad (Rocket Routine OS Launchpad): The self-serve entry point that delivers a working scorecard, Control Tower, Constitution Lite, starter Role Contract, a first real (non-demo) work result, and a running Weekly Operating Review — all on day one. Never shortened to just "Launchpad" in copy.
Leaf: Decision Impact Classification tier for routine, clearly bounded decisions; owned by AI operators within their Role Contract.
Learning: OMPRIKL artifact type created whenever the Impact phase finds a deviation; only counts as real if it changes one of the other six OMPRIKL artifacts.

M

Maturity Ladder: Four stages of operational maturity — Ad hoc, Governed, Verified, Compounding — distinct from company growth in revenue or headcount.
Metrics: OMPRIKL artifact type: how progress and current state are measured (e.g., FTT).

O

OMPRIKL: The canonical seven-artifact model (Outcomes, Metrics, Principles, Routines, Initiatives, Knowledge, Learning) every domain uses with the same structure. Always written in all caps.
Operating Modes: The three system-wide configurations — Founder, Growth, Scaling — describing how much of the company's execution runs through AI operators vs. the CEO personally.
Outcomes: OMPRIKL artifact type: verifiable results with clear boundaries, as opposed to vague goals.

P

Paperclip: Named open-source competitor: a self-hosted, agent-agnostic orchestration platform pursuing "zero-human companies," positioned as strong on execution/governance-of-config but lacking a methodology or verification layer.
Poka Yoke: Toyota Production System principle, applied to Role Contracts: design the process so an error structurally cannot occur, instead of relying on human oversight to catch it after the fact.
Principles: OMPRIKL artifact type: operational decision rules that actively guide choices (e.g., "data beats opinions"), not passive wall-values.

R

Role Contract: The persistent governance artifact that provisions an AI operator, comprising eight mandatory components (purpose/scope, outcomes/metrics, routine ownership, decision rights, tool access, quality confirmation duties, interfaces, upgrade rules). Not a prompt. Capitalized always.
Root: Decision Impact Classification tier for existential decisions (business model, financing, ownership); owned by the CEO and/or owners only.
Routines: OMPRIKL artifact type: standardized workflows with defined input, output, cadence, and quality criteria.

S

Scaling Mode: Operating Mode in which the system runs autonomously inside explicit governance and the CEO works purely at the strategic/configuration layer.
Shadow: Lowest Adoption Level: the AI operator drafts; humans execute everything that ships. Used to calibrate whether drafts approach standard.

T

Trunk: Decision Impact Classification tier for structural decisions (org design, major investment, market entry); owned by the senior team.

V

Verification-first: The governance obligation that nothing ships without a passed quality check against defined criteria; trust is verification's output, not its precondition.

W

Weekly Operating Review: The core weekly cadence routine: a fixed three-item review (what moved a metric, what's stuck, which decision is due) run before any prioritization work, active from week one of the Launchpad.