Project 28: Real Agent Business Case Study Dossier
Build a structured evidence base of successful agent businesses and one failure case.
Quick Reference
| Attribute | Value |
|---|---|
| Difficulty | Level 2: Intermediate |
| Time Estimate | 8-14 hours |
| Language | Markdown + SQL |
| Prerequisites | Product and economics basics |
| Key Topics | case-study rigor, monetization analysis, cost inference, failure analysis |
Learning Objectives
- Extract architecture and monetization patterns from public evidence.
- Distinguish facts from inferences and assumptions.
- Compare cost structures across agent business types.
- Analyze one failure case for governance and product lessons.
The Core Question You’re Answering
“What operational and business patterns separate scalable agent companies from fragile ones?”
Concepts You Must Understand First
| Concept | Why It Matters | Where to Learn |
|---|---|---|
| Evidence grading | prevents overclaiming | research methods |
| Unit economics | links architecture to margin | startup finance basics |
| Business model mapping | compares monetization durability | strategy frameworks |
| Failure decomposition | avoids survivorship bias | postmortem practice |
Theoretical Foundation
Public Source -> Structured Case Record -> Comparative Matrix -> Strategic Conclusions
Good case studies separate hard facts from reasoned inference.
Project Specification
What You’ll Build
A dossier with 5 successful cases and 1 failure case, each containing:
- Architecture snapshot
- Monetization model
- Cost structure (fact + inference labels)
- Key strategic lessons
Functional Requirements
- Common case schema
- Fact/inference tagging
- Comparative economics table
- Failure case root-cause analysis
Non-Functional Requirements
- Source traceability
- Consistent scoring rubric
- Reproducible conclusions under sensitivity tests
Real World Outcome
$ python p28_case_dossier.py --output reports/agent_business_dossier.md
[cases] success=5 failure=1
[tagging] fact=112 inference=46 assumption=18
[economics] stress_test=complete
[artifact] reports/agent_business_dossier.md
Architecture Overview
Source Collector -> Normalizer -> Case Schema Store -> Comparative Analyzer -> Dossier Writer
Implementation Guide
Phase 1: Schema and Sources
- Define case template and collect official references.
Phase 2: Comparative Analysis
- Fill architecture/monetization/cost fields and score confidence.
Phase 3: Failure and Sensitivity
- Add failure-case teardown and assumption stress tests.
Testing Strategy
- Source-link validation
- Inter-rater consistency checks
- Sensitivity analysis reruns
Common Pitfalls & Debugging
| Pitfall | Symptom | Fix |
|---|---|---|
| Marketing-only analysis | shallow conclusions | enforce architecture and cost fields |
| Missing uncertainty labels | overconfident recommendations | tag fact/inference/assumption |
| Survivorship bias | unrealistic optimism | include at least one failure case |
Interview Questions They’ll Ask
- How do you build a case study without overfitting to one company?
- How do you infer costs from incomplete public data?
- Which monetization models are robust to token volatility?
- What lessons do failure cases reveal about governance?
Hints in Layers
- Hint 1: Lock schema before collecting evidence.
- Hint 2: Quote only official disclosures for hard facts.
- Hint 3: Add scenario stress tests for margin assumptions.
- Hint 4: Keep failure analysis as detailed as success analysis.
Submission / Completion Criteria
Minimum Completion
- 5 successful + 1 failure case with source links
Full Completion
- Comparative matrix with fact/inference tags
Excellence
- Strategic memo with robust, sensitivity-tested recommendations