# Operate and Improve a Process

Status: draft · v0.1.0 · Operating System
draft reference — not a fitted standard

Keep a process controlled, useful, measurable, and adaptable after design and implementation end.

## Purpose
Give an operating owner a repeatable way to run the process, respond to exceptions, interpret evidence, and make controlled improvements without eroding the mechanism.

## Steps
1. Establish operating control — Confirm that the process can be run and governed from one authoritative definition.
2. Run and observe the process — Produce the intended work while creating proportionate evidence about how the mechanism behaves.
3. Triage and resolve exceptions — Restore flow safely while preserving information needed to remove recurring causes.
4. Review process performance — Interpret the measure set as a system and decide where attention is warranted.
5. Design and authorize improvement — Change the smallest appropriate part of the mechanism without creating hidden inconsistency.
6. Implement, verify, and learn — Close the feedback loop through controlled action and versioned learning.

## Connections
- upstream: Implement a Business System (/processes/implement-business-system) — trigger: Implementation meets stabilization and operational-acceptance criteria.
- upstream: Redesign Work and Workflow (/processes/redesign-work-and-workflow) — trigger: The redesigned workflow meets validation and transition criteria.
- downstream: Redesign Work and Workflow (/processes/redesign-work-and-workflow) — trigger: Review determines the mechanism requires material redesign.
- downstream: Implement a Business System (/processes/implement-business-system) — trigger: An approved improvement requires material configuration, integration, migration, or automation.

## Measures
- Process outcome health: The process-specific outcome measure interpreted against baseline, context, demand, and other material influences.
- Output quality: Accepted outputs without material correction divided by all outputs, supplemented by defect severity.
- Flow time: Elapsed time distribution from entry to exit, with wait and active time separated where useful.
- Controlled process use: Eligible cases following the authoritative version divided by all eligible cases.
- Exception pattern: Exceptions by type, severity, cause status, process step, interface, owner, and age per unit of demand.
- System guardrails: Context-specific guardrail set reviewed alongside outcome and flow measures.