Tutorial: Creating and Managing Cases

This tutorial walks through creating a case, progressing through stages, and managing the artifact pipeline.

1. Case Creation

Example: Research Case

Scenario: Evaluate ML framework performance for recommendation systems.

Step 1: Define Desired Outcome

1

Desired Outcome: "Comparative analysis of 3 ML frameworks with deployment recommendation"

Step 2: Select Template

1
2

Template: research_case_pattern
Stages: Literature Review → Hypothesis → Experiment → Analysis → Synthesis

Step 3: Configure Case

1
2
3
4
5
6
7
8
9
10
11
12

case:
  id: case-ml-framework-eval
  type: research
  desired_outcome: "Framework comparison with recommendation"
  owner: principal-investigator
  members:
    - human: researcher-001
    - ai: literature-summarizer-g3
    - ai: data-analyzer-g3
  pm_agent:
    knowledge_base: ml-frameworks-context
    assignment_rules: capability-based

Step 4: Activate Initial Stage

1
2
3
4
5

Stage: Literature Review
Entry Sentry: Case created (automatic)
Tasks:
  - [Mandatory] Search academic databases
  - [Discretionary] Consult domain experts (if gaps found)

2. Stage Progression

Literature Review → Hypothesis Formation

Exit Sentry Evaluation:

1
2
3
4
5

Conditions:
  - All search tasks complete: ✓
  - Summary document created: ✓
  - Peer review passed: ✓
Result: Sentry satisfied → Stage completes

Artifact Generation:

1
2
3
4
5
6
7
8

Cognitive Artifacts:
  - Search notes (ephemeral)
  - Paper annotations (ephemeral)

Synthesis Flow:
  → Intellectual Artifact: "Literature Review Summary"
  → TransitionToken: tt-lit-review-001
  → Logged to IFL

Next Stage Activation:

1
2
3

Stage: Hypothesis Formation
Entry Sentry: Literature review complete
PM-Agent Action: Assigns "Draft hypothesis" to researcher-001

3. Discretionary Task Activation

Scenario: Initial Results Inconclusive

Sentry Evaluation:

1
2
3
4
5
6
7
8
9

Stage: Experiment Execution
Discretionary Task: "Additional data collection"
Activation Condition: "Initial sample size < statistical threshold"

PM-Agent Evaluation:
  - Check: Sample size = 50
  - Threshold: 100
  - Decision: Activate discretionary task
  - Assignment: data-collector-g3

Artifact Update:

1
2
3

Cognitive Artifact: "Additional experiment data"
  → Appends to existing CaseFile
  → New TransitionToken: tt-exp-data-002

4. Artifact Pipeline Progression

Cognitive → Intellectual

Example: Experiment Notes → Analysis Report

1
2
3
4
5
6
7
8

Input: Cognitive Artifacts (raw data, observations)
Process: Synthesis Flow
  1. PM-Agent assigns "Analyze data" to data-analyzer-g3
  2. AI generates statistical analysis
  3. Human reviews and refines
  4. Peer validation
Output: Intellectual Artifact "Analysis Report"
TransitionToken: tt-analysis-001

Intellectual → Product

Example: Analysis Report → Presentation

1
2
3
4
5
6
7

Input: Intellectual Artifact (analysis)
Process: Refinement Flow
  1. PM-Agent assigns "Create presentation" to researcher-001
  2. Human synthesizes findings
  3. Stakeholder review
Output: Information Product "Framework Comparison Presentation"
TransitionToken: tt-presentation-001

Product → Capital

Example: Presentation → Methodology

1
2
3
4
5
6
7
8
9

Input: Information Product (presentation)
Process: Capitalization Flow
  1. Identify reusable components
  2. Extract evaluation methodology
  3. Human Sovereign approval (SovereignGate policy)
  4. Add to organizational knowledge base
Output: Intellectual Capital "ML Framework Evaluation Methodology"
TransitionToken: tt-capital-001
Exchange Rate: 40 hours → $8,000 capital value

5. Case Completion

Final Stage: Synthesis

Tasks:

1
2
3

- [Mandatory] Compile final recommendations
- [Mandatory] Present to stakeholders
- [Discretionary] Create deployment guide (if recommendation approved)

Exit Sentry:

1
2
3
4
5

Conditions:
  - Recommendations documented: ✓
  - Stakeholder approval: ✓
  - All artifacts in pipeline: ✓
Result: Case completes

Case Closure:

1
2
3
4
5
6
7
8
9

Status: Completed
Duration: 6 weeks
Artifacts Generated:
  - Cognitive: 15
  - Intellectual: 8
  - Product: 3
  - Capital: 1
Capital Value: $8,000
Reuse Potential: High (methodology applicable to future evaluations)

6. PM-Agent Insights

Task Assignment Log

1
2
3
4
5
6
7
8
9
10

Case: case-ml-framework-eval
PM-Agent Decisions:
  1. Assigned "Literature search" to literature-summarizer-g3
     Rationale: AI capability match, routine task
  2. Assigned "Hypothesis draft" to researcher-001
     Rationale: Requires human judgment
  3. Activated discretionary "Additional data" task
     Rationale: Sample size below threshold
  4. Assigned "Final presentation" to researcher-001
     Rationale: Stakeholder-facing, human required

7. Troubleshooting

Issue: Sentry Not Activating

Problem: Stage won’t progress despite tasks complete

Solution:

  1. Check sentry conditions explicitly
  2. Verify all artifacts generated
  3. Review PM-Agent logs for evaluation errors
  4. Manual override if justified (logged to IFL)

Issue: Artifact Pipeline Broken

Problem: Trying to create Product without Intellectual stage

Solution:

  1. Enforce “NoTeleportation” policy
  2. Create missing Intellectual Artifact
  3. Generate required TransitionToken
  4. Then proceed to Product stage
Last Updated: January 2026 Version: 1.0.0