CONCEPTS

Concepts for designing judgment in human-AI systems.

Insynergy does not treat concepts as language for thought leadership. Concepts are design units for executives, IT, governance, audit, legal, and operational teams to work on the same decision structure.

The comprehensive concept is Decision Design™. Its core mechanisms are Decision Boundary™ and Decision Log / Decision Ledger.

The definition is based on Ryoji Morii, Decision Design as Judgment Architecture: Structuring Authority and Accountability in Human-AI Systems, Insynergy Working Paper No.001, SSRN, DOI: 10.2139/ssrn.6341998. The research program now continues with The Authority Allocation Problem in AI-Augmented Organizations, Insynergy Working Paper No.002, SSRN Abstract ID: 6980098.

Insynergy Working Paper No.003 · September 2026

Reconstructing Judgment Authority from Organizational Records: Conditional Substitution and Attribution in Conceptual Omission Tests

WP-003 examines which combinations of organizational records support reconstructing judgment authority in an internal, retrospective review. Through 20 constructed cases comprising 23 runs, it explores conditional substitution between recorded information and linked role histories, and the limits of attributing a contribution to an individual information element. These conceptual observations do not establish a universal logging schema or validate accountability, audit, or institutional memory.

Read WP-003 on SSRN ↗

TOPIC CLUSTERS

Explore the Decision Design knowledge system

pillar

What Is Decision Design?

Decision Design is Ryoji Morii's judgment architecture framework for deliberately structuring authority, boundaries, escalation, and accountability in human-AI systems.

topic

Agentic AI Governance

Agentic AI governance designs legitimate authority, constraints, escalation, and accountability for AI systems that can act and transact.

topic

AI Governance Beyond Compliance

AI governance becomes operational when policy, risk, oversight, and regulation are connected to explicit decision authority and accountable records.

topic

Decision Design for Enterprise AI

Enterprise AI requires judgment architecture wherever AI participates in finance, procurement, legal, cyber, manufacturing, service, or other consequential operations.

practice

Decision Design Practice

Decision Design practice converts ambiguous judgment flows into explicit authority maps, boundaries, escalation rules, Decision Logs, and reviewable operating artifacts.

CONCEPT STRUCTURE

Decision Design™ is the comprehensive concept.

Decision Design™ is not a tool, checklist, or approval workflow. It is the architecture of judgment authority: who may decide, when AI may be delegated authority, when authority must return, and what must be recorded so responsibility remains explainable.

COMPREHENSIVE CONCEPT

Decision Design™

The design of the whole judgment architecture, including authority allocation, delegation, recovery, accountability, and decision records.

CORE CONCEPT 01

Decision Boundary™

The explicit boundary that determines how far AI may proceed, where human judgment must return, and when authority must escalate, stop, or be overridden.

It defines delegation, escalation, override, stopping conditions, and the point at which accountability becomes human or organizational.

CORE CONCEPT 02

Decision Log / Decision Ledger

The record architecture that captures how judgment authority was exercised, delegated, recovered, modified, or challenged.

It supports traceability, explainability, auditability, and the ability to revisit decisions with institutional memory.

Why judgment is beginning to fail.

AI increases recommendations, drafts, predictions, rankings, and automated actions. It also increases the number of moments where someone must accept, reject, modify, pause, or escalate a result. Organizations have increased output, but they have not increased clarity about who decides.

SSRN WORKING PAPER

Decision Design™ is judgment architecture.

The paper does not ask only how AI should be monitored. It asks how organizations should structure the boundary between machine judgment and human or organizational authority, and how that boundary should move under defined conditions.

Central question

The issue is not only what AI can do. The core question is where machine judgment ends and where human or organizational authority begins.

Design object

Decision Design™ treats authority as something to be allocated, delegated, recovered, recorded, and governed across roles, risks, and contexts.

Missing layer

Human-in-the-Loop can place an intervention point, but it does not fully define who has authority, when boundaries move, or what must be recorded.

Practical purpose

The goal is to preserve speed, accountability, auditability, and legitimate authority at the same time.

NEW PUBLICATION

The Authority Allocation Problem in AI-Augmented Organizations

Insynergy Working Paper No.002 extends the Decision Design research program from the boundary between machine and human judgment to the broader allocation of authority in AI-augmented organizations.

Source: SSRN Abstract ID: 6980098

CLARIFICATION

Decision Design™ is not a rewording of adjacent fields.

Choice Architecture, AI Alignment, and Decision Intelligence are important neighboring domains. They operate at different layers. Decision Design™ addresses the institutional layer: authority, responsibility, boundary movement, and record design before the decision is made.

Choice Architecture

Individual choice environment

Designs how options are presented once a decision-maker already exists.

AI Alignment

AI system goals and behavior

Focuses on aligning model behavior with human intent, values, or safety requirements.

Decision Intelligence

Analytics and decision support

Uses data, models, and analysis to improve decision quality and decision processes.

Decision Design™

Institutional authority and responsibility

Designs who has judgment authority, when AI may act, when authority returns, and what evidence must remain.

Decision Boundary™

Decision Boundary™ is where judgment and responsibility connect. If the boundary is unclear, the organization may move faster, but it becomes harder to explain who decided, under which authority, and why.

Symptoms:

Approvals pass, but the actual decision-maker is unclear.
An AI recommendation gradually becomes treated as a decision.
Failure cannot be attributed to any accountable authority.

Decision Log / Decision Ledger

Records do not exist only for later explanation. They become the primary material for future judgment. A durable decision record shows what was considered, what was rejected, where the boundary was drawn, and who accepted responsibility.

Who decided, who requested the decision, and who is affected

Evidence used, evidence rejected, and assumptions held open

AI recommendations and human modifications

Stopping conditions, review conditions, and exception handling

FROM CONCEPT TO PRACTICE

Turning concepts into usable operating structures.

Decision Boundary™

Where does responsible human judgment begin?

Boundary map / authority handoff rules / escalation conditions

Decision Design™

Who decides what, under which authority, and in which format?

Judgment architecture / decision authority model

Decision Log / Ledger

What must remain explainable later?

Record schema / audit trail / reviewable decision evidence

Stopping Condition

When should AI, workflow, or delegation stop?

Stop rules / exception handling / review triggers

SCOPE

What Insynergy does and does not do.

We work on

Decision authority and accountability in AI-enabled work
Governance structures across executives, IT, audit, legal, and operations
Decision records, explainability, and review conditions
Irreversible investment, governance, and organizational decisions

We do not work on

Reselling AI tools or SaaS products
Simple prompt-writing outsourcing
AI adoption without a defined decision owner
Abstract advice without scope, outputs, and responsibility

Start with the decision that has become difficult to own.

Services About