# The Four Dimensions

Every task in the Cubit database is scored on four foundational dimensions. These dimensions are conceptually independent-each captures a distinct aspect of how work relates to AI automation.

## Overview

| Dimension                 | What It Measures              | Scale |
| ------------------------- | ----------------------------- | ----- |
| **Procedural Intensity**  | How rule-based and repeatable | 0-10  |
| **Digital Accessibility** | How software-accessible       | 0-10  |
| **Physical Embodiment**   | Motor control requirements    | 0-10  |
| **Relational Depth**      | Interpersonal engagement      | 0-10  |

The first two combine into **Structural Exposure**. The second two combine into **Human Necessity**.

## Procedural Intensity

*How rule-based, scripted, and repeatable is this work?*

### What It Captures

AI systems-including large language models-perform best on tasks with:

* Clear success criteria
* Predictable structure
* Explicit decision rules
* Repeatable patterns

Work that is inherently ambiguous, improvisational, or context-dependent offers fewer footholds for automation.

### Scoring Examples

| Score    | Characteristics                              | Examples                                         |
| -------- | -------------------------------------------- | ------------------------------------------------ |
| **1-2**  | Highly variable, creative, improvisational   | Original artistic creation, crisis improvisation |
| **3-4**  | Significant judgment, some patterns          | Strategic planning, research design              |
| **5-6**  | Semi-procedural with frequent judgment       | Case analysis, diagnostic reasoning              |
| **7-8**  | Mostly procedural with occasional exceptions | Form processing with edge cases                  |
| **9-10** | Fully scripted, rule-based, repetitive       | Data entry, routine calculations                 |

### API Access

```python
tasks = client.get_job_tasks("15-1252.00")
for task in tasks["tasks"]:
    print(f"Procedural: {task['dimensions']['procedural']}/10")
    print(f"Explanation: {task['explanations']['procedural']}")
```

***

## Digital Accessibility

*How accessible are the work's inputs and outputs to software?*

### What It Captures

Software agents can only act on work they can observe and influence. This dimension measures:

* Whether inputs come from digital systems or the physical world
* Whether outputs can be delivered digitally or require physical action
* The degree of structure in the data being processed

### Scoring Examples

| Score    | Characteristics                                | Examples                                 |
| -------- | ---------------------------------------------- | ---------------------------------------- |
| **1-2**  | Analog inputs, physical environment            | Sculpture, wilderness navigation         |
| **3-4**  | Mixed analog/digital, unstructured data        | Field research, physical inspection      |
| **5-6**  | Digital with significant unstructured elements | Document analysis, media creation        |
| **7-8**  | Primarily digital, semi-structured             | Email processing, report generation      |
| **9-10** | Fully digital, structured, API-accessible      | Database queries, automated transactions |

### Key Insight

A task can be analog without being physical-reading a printed contract requires no motor skill but isn't digitally accessible. A task can be procedural without being digital-assembly line work follows strict procedures but operates in the physical world.

***

## Physical Embodiment

*Does this work require motor control, dexterity, or physical presence?*

### What It Captures

Robotics and embodied AI lag significantly behind language and reasoning capabilities. Tasks requiring:

* Fine motor manipulation
* Navigation of unstructured environments
* Complex physical coordination
* Real-time physical response

...remain protected by hardware limitations, not software limitations.

### Scoring Examples

| Score    | Characteristics                          | Examples                                        |
| -------- | ---------------------------------------- | ----------------------------------------------- |
| **1-2**  | Purely cognitive, desk-based             | Financial analysis, software design             |
| **3-4**  | Minimal physical, controlled environment | Laboratory work, equipment monitoring           |
| **5-6**  | Moderate object handling                 | Patient care, retail assistance                 |
| **7-8**  | Significant dexterity, semi-structured   | Surgical assistance, mechanical repair          |
| **9-10** | Complex motor control, unstructured      | Surgery, professional athletics, skilled trades |

### Why This Matters

A job might score high on Structural Exposure (procedural and digital) but still be protected if it requires physical presence that current robotics cannot match.

***

## Relational Depth

*How much interpersonal engagement does this work require?*

### What It Captures

Many tasks could technically be performed by AI, but their value is inseparable from human delivery:

* Patients want human doctors for empathy and trust
* Students want human teachers for mentorship
* Clients want human advisors for accountability

This dimension measures the **social contract** embedded in the work.

### Scoring Examples

| Score    | Characteristics                      | Examples                                  |
| -------- | ------------------------------------ | ----------------------------------------- |
| **1-2**  | Solitary, no interpersonal component | Solo data analysis, individual research   |
| **3-4**  | Transactional interactions only      | Simple customer transactions              |
| **5-6**  | Collaborative coordination           | Team project work, meeting facilitation   |
| **7-8**  | Persuasion, influence, teaching      | Sales, management, instruction            |
| **9-10** | Deep empathy, trust, emotional labor | Therapy, crisis counseling, pastoral care |

### Key Insight

High Relational Depth doesn't mean the work is "better" or more important. It means the work has built-in protection from displacement because humans prefer receiving it from other humans.

***

## How Dimensions Combine

### Structural Exposure (0-1 scale)

```
Structural Exposure = (Procedural + Digital) / 20
```

Identifies work that is **accessible** to software agents.

### Human Necessity (0-1 scale)

```
Human Necessity = (Physical + Relational) / 20
```

Identifies work that **requires** human presence.

### Independence

The four dimensions are designed to be **orthogonal**-knowing one score tells you little about the others:

| Task         | Procedural | Digital | Physical | Relational |
| ------------ | ---------- | ------- | -------- | ---------- |
| Data entry   | 9          | 10      | 1        | 2          |
| Surgery      | 6          | 4       | 9        | 7          |
| Therapy      | 3          | 3       | 2        | 10         |
| Field repair | 5          | 3       | 8        | 4          |

This independence is what enables nuanced analysis-each dimension contributes distinct information.

***

## Accessing Dimension Data

### Task-Level (Professional+)

```python
tasks = client.get_job_tasks("29-1141.00")  # Registered Nurses

for task in tasks["tasks"][:3]:
    print(f"\nTask: {task['task'][:60]}...")
    print(f"  Procedural: {task['dimensions']['procedural']}")
    print(f"  Digital: {task['dimensions']['digitization']}")
    print(f"  Physical: {task['dimensions']['physicality']}")
    print(f"  Relational: {task['dimensions']['socio_emotional']}")
```

### With Explanations

```python
tasks = client.get_job_tasks("29-1141.00", include_explanations=True)

task = tasks["tasks"][0]
print(task["explanations"]["procedural"])
# "Involves established clinical protocols, but requires judgment 
#  for patient-specific adaptations."
```

***

## Next Steps

* [Understanding Scores](/core-concepts/understanding-scores.md) - How dimensions roll up to job scores
* [Quadrant Classification](/core-concepts/quadrants.md) - The four strategic zones
* [Task-Level Analysis Guide](/guides/task-analysis.md) - Working with dimension data


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.maidenlabs.tools/core-concepts/four-dimensions.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.