# Real-World Examples

Production scenarios demonstrating VAPORA solving business problems.

## Use Cases

### 01: Code Review Pipeline
**File**: `01-code-review-workflow.rs`

Automated multi-agent code review with cost optimization.

**Business Value**:
- Review 50 PRs/day consistently
- 99% accuracy matching human review
- $500+/month savings vs manual review
- Instant feedback to developers

**Architecture**:
1. Ollama (free) - Static analysis
2. GPT-4 ($10/1M) - Quality review
3. Claude ($15/1M) - Architecture review

**Run**:
```bash
cargo run --example 01-code-review-workflow
```

### 02: Documentation Generation
**File**: `02-documentation-generation.rs`

Automatically generate and maintain API documentation.

**Business Value**:
- Docs always in sync with code
- 2-3 week lag → instant updates
- 99%+ accuracy vs manual docs
- $1000+/month savings

**Architecture**:
1. Ollama - Code analysis
2. Claude - Doc writing
3. GPT-4 - Quality check

**Run**:
```bash
cargo run --example 02-documentation-generation
```

### 03: Issue Triage
**File**: `03-issue-triage.rs`

AI-powered issue classification and routing.

**Business Value**:
- 200+ issues/month triaged instantly
- Consistent classification criteria
- 95% accuracy
- $1000+/month savings (vs 20 hours manual work)

**Architecture**:
1. Ollama (free) - Initial classification
2. Claude ($0.08) - Detailed analysis when needed

**Run**:
```bash
cargo run --example 03-issue-triage
```

## Cost Analysis Pattern

Real-world examples use **cost-aware multi-stage pipelines**:

```
Stage 1 (Cheap): Ollama/GPT-4 for routing
  ↓
Stage 2 (Premium): Claude only when needed
  ↓
Stage 3 (Validation): GPT-4 for quality check
```

This achieves:
- 95%+ accuracy at 10% of premium cost
- Fast iteration with feedback
- Controlled budget

## Key Metrics

Each example demonstrates:

| Metric | Typical Value |
|--------|--------------|
| Accuracy | 95%+ vs manual |
| Speed | 10-100x faster |
| Cost | 1-10% of manual |
| Coverage | 100% automated |

## Adapting for Your Use Case

1. **Choose pipeline stages**: Cheap → Medium → Premium
2. **Adjust triggers**: When to escalate to next stage
3. **Set budgets**: Per-role monthly limits
4. **Monitor quality**: Track accuracy vs manual baseline

## Next Steps

- Implement case-specific routing rules
- Integrate with your systems (GitHub, Jira, etc.)
- Monitor and adjust thresholds
- Expand to additional use cases

---

Navigate to `../` for more examples.