# org-designs — Organizational Theory for the AI-Native Firm

A working collection on how organizations coordinate knowledge work, and how those classical
foundations extend into **AI-augmented / AI-native** organizational design. It has grown from a
single research synthesis into a small program with **four layers**:

1. **Research** — the seminal literature, its synthesis into a design framework, and derivative analyses.
2. **Courses** — two interactive, browser-based courses built on that research.
3. **The Configurator** — a diagnostic instrument that maps a unit's work onto AI-organization configurations.
4. **`.ai/`** — the working set: assessments, design strategy, and implementation plans.

> **Layout note.** The planned Tier 2 reorganization has been executed after the configurator v0.2
> rebuild. The implementation record lives at
> [`.ai/tasks/03_org_designs_reorg_plan.md`](./.ai/tasks/03_org_designs_reorg_plan.md).

---

## 1. Research

| File | What it is |
| --- | --- |
| [`research/collaboration-bottlenecks-knowledge-work.pdf`](./research/collaboration-bottlenecks-knowledge-work.pdf) | **The anchor review.** Seminal organizational research organized into 10 categories of collaboration bottlenecks; 31 numbered references. The raw input. |
| [`research/org-theory-synthesis.md`](./research/org-theory-synthesis.md) | **The synthesis.** Reframes the anchor review into a *design framework* — a unifying thesis, twelve intellectual lineages, a proposed set of orthogonal design axes, live debates, and the modern gaps (distributed work, AI). Hyperlinked, with a full DOI reference list. |

The synthesis also draws on companion research in the sibling `../human-org-role/` folder:
`dyadic-collaboration-synthesis.md`, and the working papers `03_cognitive_electricity_report.md`
and `ai_first_principles_research.md`.

> **Consolidation note.** Three earlier drafts (`01_org_architecture_components.md`, `02_`, and
> `03_org-theory-synthesis-linked.md`) were merged into the canonical `research/org-theory-synthesis.md`.

## 2. Courses

Two distinct interactive courses, both single-file, zero-dependency HTML — open in any browser.

### 2a. Designing the Organization (`courses/`)

The information-processing → macro-axes → dyad → AI-native arc.

| File | What it is |
| --- | --- |
| [`courses/org-design/index.html`](./courses/org-design/index.html) | **Canonical.** The v2 "evidence & executive edition": adds an executive on-ramp, a measurement layer, named case vignettes, objection-handling, and embeds the configurator as **Part 5 — Diagnostic**. |
| [`courses/org-design/archive/v1.html`](./courses/org-design/archive/v1.html) | The original v1 (four-part walkthrough). Superseded by v2; kept as history. |
| [`courses/org-design/archive/v1.ko.html`](./courses/org-design/archive/v1.ko.html) | Korean v1 translation, archived with v1. A full Korean v2 translation is still a separate content pass; this reorg does not relabel the stale v1 translation as current. |

The build history of the v1 → richer-course augmentation is recorded in
[`.ai/tasks/02_course_augmentation_plan.md`](./.ai/tasks/02_course_augmentation_plan.md).

### 2b. Authority by Design (`courses/authority-by-design/`)

A second, emerging course track on the economics of authority — the Aghion–Tirole formal/real
authority lineage translated into an executive decision-architecture program.

| File | What it is |
| --- | --- |
| [`.../01_aghion_tirole_lineage.md`](./courses/authority-by-design/notes/01_aghion_tirole_lineage.md) | The Aghion–Tirole benchmark and the research programs it spawned. |
| [`.../02_core_questions.md`](./courses/authority-by-design/notes/02_core_questions.md) | A seminar-ready taxonomy of the authority-allocation questions. |
| [`.../03_executive_course_outline.md`](./courses/authority-by-design/notes/03_executive_course_outline.md) | The proposed executive course architecture ("Authority by Design"). |
| [`academic-lineage.html`](./courses/authority-by-design/academic-lineage.html) | Interactive academic-lineage explorer (built artifact). |

## 3. The Configurator (`configurator/`)

A diagnostic instrument: twenty-two questions about a unit's **work, constraints, current operating
baseline, and intent** — never its structural preferences — returning target, feasible-now, C1-C6
candidate rank, calibrated band, distance gap, tensions, exclusions, influences, and roadmap.
Embedded as Part 5 of the org-design course.

| File | What it is |
| --- | --- |
| [`configurator/ai-org-configurator.html`](./configurator/ai-org-configurator.html) | The interactive v0.2 instrument. |
| [`configurator/ai-org-configurator.schema.js`](./configurator/ai-org-configurator.schema.js) | Canonical 22-question schema, dimensions, presets, caps, and validation. |
| [`configurator/ai-org-configurator.engine.js`](./configurator/ai-org-configurator.engine.js) | Pure target, cap, feasible-now, candidate, tension, influence, and envelope engine. |
| [`configurator/ai-org-configurator.calibration.json`](./configurator/ai-org-configurator.calibration.json) | Deterministic, generated calibration artifact. |
| [`configurator/ai-org-configurator.app.js`](./configurator/ai-org-configurator.app.js) | Browser state, rendering, strict v2 permalink, export, reset, and report preview. |
| [`configurator/ai-org-configurator.md`](./configurator/ai-org-configurator.md) | Generated v0.2 design specification and worked example. |

The v0.2 rebuild corrected scoring, added measured governance, restored C5, separates target from
feasible-now, uses rank/band/gap instead of primary percentages, and is covered by Node and
Playwright tests.

## 4. `.ai/` — working set

| File | What it is |
| --- | --- |
| [`.ai/analyses/01_org_design_framework_and_configurator_assessment.md`](./.ai/analyses/01_org_design_framework_and_configurator_assessment.md) | Rigorous assessment of the configurator and the course framework (findings, verification, gaps). |
| [`.ai/analyses/02_configurator_v02_improvement_strategy.md`](./.ai/analyses/02_configurator_v02_improvement_strategy.md) | Improvement strategy for the questionnaire and scoring method (v0.1 → v0.2). |
| [`.ai/analyses/03_feature_coordination_modes_matrix_analysis.md`](./.ai/analyses/03_feature_coordination_modes_matrix_analysis.md) | The Top-Down / Co-Pilot / Autopilot × 8-department coordination-mode matrix. |
| [`.ai/tasks/01_configurator_v02_implementation_plan.md`](./.ai/tasks/01_configurator_v02_implementation_plan.md) | Execution record and TDD plan for the v0.2 configurator rebuild. |
| [`.ai/tasks/02_course_augmentation_plan.md`](./.ai/tasks/02_course_augmentation_plan.md) | The plan behind the v1 → richer-course augmentation (relocated here from `courses/`). |
| [`.ai/tasks/03_org_designs_reorg_plan.md`](./.ai/tasks/03_org_designs_reorg_plan.md) | This directory's reorganization plan and Tier 2 execution record. |

---

## The unifying thesis

Organizational theory converges on a single deep idea:

> An organization is an **information-processing and coordination system**, and its structure is a
> set of **design choices** that trade off the cost of processing information against the
> **uncertainty and interdependence** of its work.

The lineage runs from Simon's bounded rationality (1947) → March & Simon / Cyert & March
(behavioral theory of the firm) → Galbraith's information-processing design view (1973–74) → modern
network, modularity, and self-management research — and now into the AI-native firm as a cybernetic
control system (operating loop → meta-loop, oriented by intent, bounded by governance).

### The two research artifacts and how they relate

The **PDF** is the raw input — a literature review that catalogs *where knowledge work breaks down*.
The **synthesis** is the output — it reorganizes the same body of thinkers from "categories of
bottlenecks" into "orthogonal axes of design choice," then extends them toward modern distributed
and AI-augmented forms. Nearly every author in the PDF reappears in the synthesis, recast from
*what breaks* into *what to decide when designing an organization*.

**The bottleneck taxonomy (PDF)** — ten categories, each anchored to seminal thinkers:

1. **Goal & priority** — Locke & Latham; Drucker (MBO); Kaplan & Norton (Balanced Scorecard).
2. **Decision-rights & governance** — Jensen & Meckling (agency); Aghion & Tirole (formal vs. real authority); Mintzberg.
3. **Information & context** — Galbraith; Walsh & Ungson (organizational memory); Nonaka & Takeuchi (SECI); Allen; Wegner (transactive memory).
4. **Work-structuring (interfaces & dependencies)** — Thompson (interdependence types); Malone & Crowston (coordination theory); Conway's Law.
5. **Process & throughput** — Goldratt (Theory of Constraints); Hammer & Champy (reengineering); Womack et al. (Lean).
6. **Incentive & performance-system** — Kerr ("rewarding A while hoping for B"); Deci & Ryan (self-determination); Herzberg (two-factor).
7. **Tooling & workflow** — Grudin (groupware); Suchman (situated action); Orlikowski (duality of technology).
8. **Communication & sensemaking** — Weick (sensemaking); Daft & Lengel (media richness); Clark & Brennan (grounding); Allen.
9. **Trust, psychological safety & conflict** — Edmondson (psych safety); Mayer et al. (trust); Jehn (task vs. relationship conflict).
10. **Leadership attention & span-of-control** — Mintzberg (managerial work); Graicunas / Urwick (span of control); Simon (attention scarcity).

**The design-framework synthesis (Markdown).** Twelve theory clusters (A–L) resolve into a
**four-spine core** — (1) decision rights, (2) coordination mechanism, (3) channel/network design,
(4) knowledge/memory architecture — plus three genuinely orthogonal modifiers: **boundary
permeability, incentive alignment, and psychological-safety climate.** Correlated dimensions
(formalization, span of control, mechanistic–organic) collapse into a single "structuring
intensity" factor; task interdependence and coupling are treated as *contingency inputs* the design
must fit, not free choices. The gaps where the classical canon is thin — and where new work should
focus — are **distributed/remote work, platform/ecosystem organizations, radical self-management,
and AI-augmented cognition.**

The synthesis (`research/org-theory-synthesis.md`) is structured as: **TL;DR** · **Key Findings** ·
**Details** (I. intellectual landscape & lineages · II. validating the three proposed axes ·
III. a fuller orthogonal-axis set · IV. debates, tensions, gaps) · **Recommendations** ·
**Caveats** · **References**.

---

## How to use this material

- **Fast orientation:** read the TL;DR and Key Findings of [`research/org-theory-synthesis.md`](./research/org-theory-synthesis.md).
- **Teaching / learning:** open [`courses/org-design/index.html`](./courses/org-design/index.html) — the canonical course, ending in the embedded diagnostic.
- **Diagnose a unit:** open [`configurator/ai-org-configurator.html`](./configurator/ai-org-configurator.html) directly, or reach it as Part 5 of the course.
- **Authority track:** start with [`.../03_executives_course_outline.md`](./courses/authority-by-design/notes/03_executive_course_outline.md).
- **Citation:** use the References section of the synthesis (DOIs and verified open-access links). Note the caveats: practitioner works (Laloux, Robertson, Skelton & Pais, Hamel) are influential but not peer-reviewed, and the orthogonality claims are conceptual and await empirical testing.