Human Decision Is Not Optional

A named person accepts, rejects, escalates, or returns the recommendation. That act is the decision. Until that act is recorded, the case is still a proposal. Authority stays with the person who can accept the operational consequence. The signed-in Decision Case keeps the person and the proposal together. The recorded human decision is mandatory. A green dashboard does not imply it. The system does not accept the recommendation on that person's behalf.

Reliability work skips the act because everything around it looks finished. The dashboard is green. The alert has been cleared. The question has been sharpened. The recommendation is already written. The meeting moves on. Later the record is cited as if someone had decided. No named person had accepted, rejected, escalated, or returned the recommendation. The case was still a proposal. Authority had not moved.

Sync keeps that split on the signed-in Decision Case. A signed-in user completes the case in a fixed order: Question, Evidence, Recommendation, Human decision, Action, Verification, and Learning. Orville Davis states that order in Field Manual v0. The manuals index lives at /manuals. This essay is why the Human decision step cannot be skipped, implied, or filled in by the system. The manual is the public contents. The chapter is the act.

That act is the decision

A named person accepts, rejects, escalates, or returns the recommendation. That act is the human decision. Accept, reject, escalate, and return are the same kind of record: a named person has acted on the recommendation, and the case holds that act. Until one of those acts is recorded, the case is still a proposal.

The Human decision chapter publishes that step. It is step 04 on the spine. It sits after the recommendation and before action. The signed-in loop keeps the act explicit: the case shows who decided, and the recommendation remains visible beside the decision.

The recorded human decision is mandatory.

Skipping the act does not leave a quieter version of the same case. It leaves a proposal. A recommendation that nobody has accepted, rejected, escalated, or returned is still waiting on the person who can accept the operational consequence.

Until the act is recorded, the case is still a proposal

Until that act is recorded, the case is still a proposal. The recommendation can be complete. The evidence can sit on the case. The wording can be ready for a meeting. None of that records the decision. The case holds a proposed next action, the evidence it uses, and the uncertainty it leaves in view. Authority stays where it already sits.

Authority stays with the person who can accept the operational consequence.

That person is the one who can accept, reject, escalate, or return. The software that drafted the proposal does not hold that authority. A status display does not hold it. Silence after the draft does not transfer it. The system does not auto-accept the recommendation. Stage-1 readiness means a signed-in user can complete the Decision Case, and completion includes this named act. It does not include a rule that treats an unreviewed proposal as accepted.

The case keeps the person beside the proposal

The signed-in Decision Case keeps the person and the proposal together. Who decided stays visible beside the recommendation. A later reader can see both on the same record: the proposal that was offered, and the named person who accepted, rejected, escalated, or returned it.

The reader does not have to infer the decision from a green tile, a cleared alert, or the fact that a recommendation exists. The chapter's sentence is specific. The case shows who decided, and the recommendation remains visible beside the decision. A record that cannot name the person has not recorded the act.

A proposal is not that act

Recommend is not authorize. A recommendation states a proposed next action, the evidence it uses, and the uncertainty it leaves in view. Drafting the proposal leaves authority where it already sits. That split is the subject of Recommend Is Not Authorize. A drafted proposal is not permission.

This essay is the next sentence. The proposal and the recorded decision are different objects. The human decision is the act that has to be on the case before the proposal is treated as decided. Evidence is required before a recommendation is treated as a decision. The named act is required as well. One does not stand in for the other. Until accept, reject, escalate, or return is recorded, the recommendation remains a proposal.

Read the two steps as separate chapters: Recommendation and Human decision. The first can be drafted. The second is the named person's act on that draft.

Inputs do not record the act

A green dashboard does not imply the decision. Dashboard Is Not Decision states that seeing plant or maintenance state — a KPI tile, a trend, a health score, a red, amber, or green status — is not a Decision Case and not a decision. A green dashboard can sit beside the case. It does not accept the recommendation.

An alert does not imply the decision either. Alert Is Not Decision states that seeing, acknowledging, or silencing an alert is not the named act. Clearing a queue does not record who decided.

A question does not imply it. Question Is Not Decision states that recording or refining a reliability or maintenance question is not the decision. The question is step 01. It is written before evidence and before a recommendation. The human decision is step 04, after a recommendation exists to decide. An open question, a clarified question, and a rephrased question are still questions.

Those inputs are not the decision. This essay is the recorded act itself.

The dashboard, the alert, and the question can all be present. The recommendation can be present. The decision is still the named person's accept, reject, escalate, or return, written on the case. Until that act is recorded, the case is still a proposal.

The recorded act does not execute the work

Accepting a recommendation records who decided. It does not write a work order, isolate equipment, or change the plant. Action Is Not Execution states the next boundary. Action is the record of the work the human decision authorized. On the case it is a disposition: work intent, held with the decision that produced it. ACTION stays locked until authorized execution systems write the work order or isolate the equipment. Sync does not perform that write. The product does not dispatch unsupervised work to equipment.

Verification Is Not Optional states the gate after that intent. Authorization records who decided. It does not record whether the action did what it was supposed to do. The signed-in Decision Case stays open until that check is on the case. The human decision is not the check. Learning keeps the closed case — who decided, what was authorized, and what was checked — so a later question can see it. Learning does not inherit an acceptance that was never recorded.

Honesty Boundary Is Not Optional states the limit that governs the act. A Decision Case that pretends plant authority Sync does not hold is not a Decision Case. Naming who decided does not give the product authority to execute the work. This essay does not claim that Sync executes plant work. It does not claim that self-guided onboarding is a live product path. The recorded decision is a named human act on a recommendation.

Read the later steps in the Action chapter and the Verification chapter. The standing rule sits in Honesty boundaries. None of them treats a missing human decision as decided, and none of them treats the recorded act as plant execution.

Where the public statement lives

Field Manual v0 is the public contents of this loop. Start at the manuals index or open Sync Field Manual directly. Human decision is step 04. It sits after the recommendation and before action. Later editions can deepen a chapter. The spine stays in this order. The recorded act stays mandatory.

What this article is not claiming

This is an essay about the Decision Case order, not a customer case study. It names no plant, states no savings figure, and claims no prevented failure. It states no OEM limit and no operating threshold. It does not claim that Sync executes plant work, writes a work order, isolates equipment, or controls a plant. It does not claim that the system auto-accepts a recommendation. It does not claim that self-guided onboarding is a live product path.

Stage-1 readiness means a signed-in user can complete the Decision Case — question, evidence, recommendation, human decision, action, verification, and learning — and Field Manual v0 describes that journey. This edition does not describe plant execute, a live connector tag pull, SMTP invite delivery, or automatic revocation of access on expiry as live. It does not describe Sync writing work orders, isolating equipment, or controlling the plant. Simulated or seeded telemetry and assets are practice records. They are not live plant results. Self-guided onboarding is not claimed as a live product path.

The series continues with Unknown Is Not Clear, on why missing evidence is recorded as unknown, not as clear or ready, and why the operator must not treat unknown as go. Companion reading: Recommend Is Not Authorize on why a drafted proposal is not permission, Honesty Boundary Is Not Optional on why Sync states what is simulated, seeded, or out of scope, Action Is Not Execution on why a recorded ACTION disposition is not plant execution, Verification Is Not Optional on why the case stays open until the check is recorded, and Question Is Not Decision on why recording a reliability or maintenance question is not the human decision. A Reliability Assessment asks whether the records can support a conclusion. A Strategic Pilot is a governed proof around one operating decision. The Human decision chapter is the public statement of the act this essay describes.