consensus/examples/02-ai-payment/README.md

Example 02: AI-Powered Payment Validation

This example demonstrates how AI consensus validates cross-chain payments before execution.

What It Shows

• AI agent analyzing payment requests
• Confidence scoring for transactions
• Fraud detection using AI
• Learning from payment outcomes

Run the Example

cd examples/02-ai-payment
go run main.go

Run the Tests

go test -v

Expected Output

=== AI Payment Validation ===

Payment Request #1:
  Amount:   1000 LUX
  From:     ethereum
  To:       lux
  
AI Analysis:
  ✓ Confidence: 0.92 (HIGH)
  ✓ Risk Level: LOW
  ✓ Decision: APPROVE
  Reasoning: Normal transaction pattern, amount within limits

Payment Request #2:
  Amount:   50000 LUX
  From:     ethereum
  To:       lux
  
AI Analysis:
  ⚠ Confidence: 0.45 (LOW)
  ⚠ Risk Level: HIGH
  ✗ Decision: REJECT
  Reasoning: Unusual amount, potential fraud

=== Learning from Outcomes ===
✓ AI updated with 2 training examples
✓ Model weights adjusted based on feedback

Key Concepts

1. AI Decision Making: Agent analyzes payment before execution
2. Confidence Scoring: Determines how sure the AI is
3. Risk Assessment: Identifies potentially fraudulent transactions
4. Continuous Learning: Improves from successful/failed payments

Integration Pattern

// 1. Create AI agent for payment validation
agent := ai.NewAgent(nodeID, model, quasar, photon)

// 2. Analyze payment request
decision, err := agent.ProposeDecision(ctx, txData, context)

// 3. Check confidence threshold
if decision.Confidence < 0.7 {
    // Reject low-confidence payments
    return errors.New("confidence too low")
}

// 4. Learn from outcome
agent.AddTrainingData(example)

Play With It

Try modifying:

• Confidence thresholds
• Risk assessment criteria
• Training data feedback
• Transaction patterns