Human: "Create a brand awareness campaign for our new sustainable sneaker line, targeting eco-conscious millennials with a $75K budget"

AI Agent: I'll create that campaign for you. Let me set up:
→ create_brand_agent({ name: "EcoStep Footwear" })
→ create_campaign({
    brandAgentId: "ba_123",
    name: "Sustainable Sneaker Launch",
    prompt: "Target eco-conscious millennials interested in sustainable fashion and footwear",
    budget: { total: 75000, currency: "USD" }
  })

✅ Campaign created! Using built-in optimization to find the best inventory mix for eco-conscious millennials. Expected reach: 2.5M people.

Human: "Find me premium CTV inventory under $35 CPM"

AI Agent: → discover_publisher_products({
    formats: ["video"],
    deliveryType: "ctv",
    maxCpm: 35,
    quality: "premium"
  })

Found 12 premium CTV options under $35 CPM:
📺 Hulu Premium CTV - $28 CPM, 85% completion rate
📺 Samsung TV Plus Premium - $32 CPM, viewability 95%
📺 Roku Premium Inventory - $30 CPM, brand-safe content
...

Human: "How is our sneaker campaign performing? Focus on CTV vs display"

AI Agent: → analyze_inventory_performance({
    campaignId: "camp_456",
    groupBy: ["format", "inventorySource"]
  })

📊 Performance Breakdown (Last 7 days):
CTV Inventory: $45K spent, 1.8M impressions, 0.85% CTR
Display Inventory: $22K spent, 3.2M impressions, 0.31% CTR

Recommendation: CTV is delivering 2.7x higher engagement. Consider shifting more budget to CTV formats.

{
  "mcpServers": {
    "scope3-campaigns": {
      "command": "npm",
      "args": ["start"],
      "cwd": "/path/to/scope3-campaign-api",
      "env": {
        "SCOPE3_API_KEY": "your-api-key-here"
      }
    }
  }
}

import mcp

# Connect to Scope3 MCP server
client = mcp.Client("http://localhost:3001/mcp")

# Execute campaign creation
result = await client.call_tool("create_campaign", {
    "brandAgentId": "ba_123",
    "name": "AI-Managed Campaign",
    "prompt": "Optimize for brand awareness among tech professionals",
    "budget": {"total": 50000, "currency": "USD"}
})

Creative Agent: Manages creative assets and A/B testing
→ create_creative(), update_creative()

Campaign Manager Agent: Handles campaign setup and goals
→ create_campaign(), update_campaign()

Optimization Agent: Manages inventory and performance
→ discover_publisher_products(), adjust_inventory_allocation()

Reporting Agent: Provides insights and analysis
→ analyze_inventory_performance(), list_campaigns()

Session 1: Create campaign "camp_123"
Session 2: "Adjust that campaign's budget" → Knows to use "camp_123"
Session 3: "How is it performing?" → Analyzes performance for "camp_123"

{
  "success": true,
  "data": {
    "campaignId": "camp_123",
    "name": "Q4 Performance Campaign",
    "status": "active",
    "budget": {
      "total": 50000,
      "spent": 12500,
      "remaining": 37500,
      "currency": "USD"
    }
  },
  "humanReadable": "✅ Campaign 'Q4 Performance Campaign' created successfully with $50K budget. Currently active with $37.5K remaining."
}

{
  "success": false,
  "error": "INSUFFICIENT_BUDGET",
  "message": "Budget allocation ($75,000) exceeds campaign total ($50,000)",
  "suggestions": [
    "Reduce individual allocations",
    "Increase campaign budget",
    "Remove some inventory options"
  ]
}

Human: "I want to spend more on premium inventory"
↓
Agent Processing: Identify campaign, determine "premium" criteria, calculate budget shift
↓
Tool Call: adjust_inventory_allocation({
  campaignId: "current_campaign",
  adjustments: [{
    inventoryOptionId: "premium_options",
    budgetIncrease: calculated_amount
  }]
})

// Raw API Response
{
  "inventoryOptions": [
    {"name": "CTV Premium", "cpm": 28, "performance": 0.85},
    {"name": "Display Standard", "cpm": 12, "performance": 0.31}
  ]
}

// AI Agent Response
"Your CTV Premium inventory is performing 2.7x better than Display (0.85% vs 0.31% CTR),
but costs 2.3x more ($28 vs $12 CPM). The efficiency is worth the premium -
consider shifting more budget to CTV."

After creating campaign:
"✅ Campaign created with built-in optimization.

Next steps:
• discover_publisher_products() - Explore specific inventory options
• create_creative() - Add creative assets
• analyze_inventory_performance() - Monitor performance after 48 hours"

Human: "Set up a complete campaign for our new product launch"

Agent Workflow:
1. create_brand_agent() - If needed
2. create_creative() - For launch assets
3. create_campaign() - Main campaign setup
4. create_inventory_option() - For specific inventory requirements
5. Provide launch checklist and monitoring plan

npm run dev

curl -X POST http://localhost:3001/mcp \
  -H "Content-Type: application/json" \
  -H "x-scope3-api-key: your-key" \
  -d '{"tool": "create_campaign", "parameters": {...}}'

// Test campaign creation
const result = await testTool("create_campaign", {
  brandAgentId: "test_brand_agent",
  name: "Test Campaign",
  prompt: "Test targeting tech enthusiasts",
  budget: { total: 10000, currency: "USD" },
});

console.log("Campaign created:", result.data.campaignId);

// Test insufficient budget error
const result = await testTool("create_inventory_option", {
  campaignId: "camp_123",
  budgetAllocation: { amount: 999999, currency: "USD" }, // Exceeds campaign budget
});

console.log("Error handled:", result.error);

const CreateCampaignParams = z.object({
  brandAgentId: z.string(),
  name: z.string().min(1).max(100),
  budget: z.object({
    total: z.number().positive(),
    currency: z.string().length(3),
  }),
});