DHL prospect scoping

{
  "code": "DHL Prospect Scoping — Technical Proposal\n\n## DHL Integration Requirements\n\n### 1. API Endpoints Required\n\n| Endpoint | Method | Purpose |\n|----------|--------|---------|\n| `/api/v1/classify` | POST | Single product HTS classification |\n| `/api/v1/classify/bulk` | POST | Batch classification (up to 500 products) |\n| `/api/v1/duty/calculate` | POST | Duty + fee calculation for classified product |\n| `/api/v1/pga/check` | POST | PGA requirements (FDA, USDA, EPA, CPSC) |\n| `/api/v1/entry-type` | POST | Entry type determination (01/11/52 etc.) |\n| `/api/v1/isf/generate` | POST | ISF 10+2 data generation for ocean |\n| `/api/v1/jobs/{id}` | GET | Poll async batch status |\n\nAll endpoints already exist in AskRosetta API (app.py + corridor_engine/).\n\n### 2. Data Formats\n\n**Input** (DHL → C8):\n```json\n{\n  \"shipment_id\": \"DHL-1234567890\",\n  \"products\": [{\n    \"sku\": \"DHL-SKU-001\",\n    \"description\": \"Men's cotton polo shirt\",\n    \"origin_country\": \"BD\",\n    \"destination_country\": \"US\",\n    \"value_usd\": 12.50,\n    \"weight_kg\": 0.3,\n    \"image_url\": \"https://...\"  // optional\n  }],\n  \"corridor\": \"BD_US\",\n  \"service_level\": \"express\"  // express|standard\n}\n```\n\n**Output** (C8 → DHL):\n```json\n{\n  \"shipment_id\": \"DHL-1234567890\",\n  \"classifications\": [{\n    \"sku\": \"DHL-SKU-001\",\n    \"hts_code\": \"6105.10.00\",\n    \"hts_description\": \"Men's or boys' shirts, knitted, of cotton\",\n    \"confidence\": 0.92,\n    \"duty_rate\": 0.199,\n    \"estimated_duty_usd\": 2.49,\n    \"mpf_usd\": 0.56,\n    \"hmf_usd\": 0.03,\n    \"total_landed_cost_usd\": 15.58,\n    \"pga_flags\": [],\n    \"section_301_applies\": false,\n    \"source\": \"rosetta_core\"\n  }],\n  \"entry_recommendation\": \"11\",  // Informal entry (<$2500)\n  \"isf_required\": false,  // Only for ocean\n  \"processing_time_ms\": 850\n}\n```\n\n### 3. Volume Estimates\n\nDHL Express handles ~5M US import shipments/year. At C8 integration:\n- **Phase 1 (pilot)**: 1,000 classifications/day (~30K/month)\n- **Phase 2 (rollout)**: 10,000/day (~300K/month)  \n- **Phase 3 (full)**: 50,000+/day (~1.5M/month)\n\nCurrent AskRosetta capacity: ~100 req/s on Cloud Run (auto-scaling).\n\n### 4. Pricing Model\n\n| Tier | Volume/month | Price/classification | Monthly |\n|------|-------------|---------------------|---------|\n| Pilot | Up to 30K | $0.15 | $4,500 |\n| Growth | 30K-300K | $0.08 | $8,000-$24,000 |\n| Enterprise | 300K+ | $0.03 | $9,000+ |\n\n**Value proposition for DHL**:\n- Current manual HTS classification: ~$2-5 per product (broker time)\n- C8 automated: $0.03-0.15 per product\n- **90-98% cost reduction** on classification\n- Reduced CBP penalties from misclassification\n- ISF automation for ocean shipments\n- PGA flag detection prevents holds at port\n\n### 5. Integration Architecture\n\n```\nDHL Systems → API Gateway → AskRosetta Cloud Run (Zurich)\n                                    ↓\n                          Rosetta Core (deterministic)\n                                    ↓ (fallback)\n                          Gemini Flash (AI classification)\n                                    ↓\n                          Results + Duty Calc → DHL\n```\n\n**Authentication**: API key (X-Api-Key header) with per-key rate limits and usage metering.\nAlready implemented in stripe_billing.py.\n\n### 6. SLA Commitments\n\n- **Latency**: <2s for single classification, <60s for 100-product batch\n- **Availability**: 99.9% (Cloud Run auto-scaling, multi-region possible)\n- **Accuracy**: >90% HTS accuracy at 6-digit level (with confidence scores)\n- **Data residency**: EU (Zurich) primary, US failover available\n\n### 7. Next Steps\n\n1. Request DHL API docs for their shipment data format\n2. Build DHL-specific adapter mapping their SKU format to ClassificationRequest\n3. Set up dedicated API key with DHL-specific rate limits\n4. Pilot with 100 real DHL shipments to calibrate accuracy\n5. Weekly accuracy review during pilot (target: 92%+ at 6-digit)",
  "content": "DHL Prospect Scoping — Technical Proposal\n\n## DHL Integration Requirements\n\n### 1. API Endpoints Required\n\n| Endpoint | Method | Purpose |\n|----------|--------|---------|\n| `/api/v1/classify` | POST | Single product HTS classification |\n| `/api/v1/classify/bulk` | POST | Batch classification (up to 500 products) |\n| `/api/v1/duty/calculate` | POST | Duty + fee calculation for classified product |\n| `/api/v1/pga/check` | POST | PGA requirements (FDA, USDA, EPA, CPSC) |\n| `/api/v1/entry-type` | POST | Entry type determination (01/11/52 etc.) |\n| `/api/v1/isf/generate` | POST | ISF 10+2 data generation for ocean |\n| `/api/v1/jobs/{id}` | GET | Poll async batch status |\n\nAll endpoints already exist in AskRosetta API (app.py + corridor_engine/).\n\n### 2. Data Formats\n\n**Input** (DHL → C8):\n```json\n{\n  \"shipment_id\": \"DHL-1234567890\",\n  \"products\": [{\n    \"sku\": \"DHL-SKU-001\",\n    \"description\": \"Men's cotton polo shirt\",\n    \"origin_country\": \"BD\",\n    \"destination_country\": \"US\",\n    \"value_usd\": 12.50,\n    \"weight_kg\": 0.3,\n    \"image_url\": \"https://...\"  // optional\n  }],\n  \"corridor\": \"BD_US\",\n  \"service_level\": \"express\"  // express|standard\n}\n```\n\n**Output** (C8 → DHL):\n```json\n{\n  \"shipment_id\": \"DHL-1234567890\",\n  \"classifications\": [{\n    \"sku\": \"DHL-SKU-001\",\n    \"hts_code\": \"6105.10.00\",\n    \"hts_description\": \"Men's or boys' shirts, knitted, of cotton\",\n    \"confidence\": 0.92,\n    \"duty_rate\": 0.199,\n    \"estimated_duty_usd\": 2.49,\n    \"mpf_usd\": 0.56,\n    \"hmf_usd\": 0.03,\n    \"total_landed_cost_usd\": 15.58,\n    \"pga_flags\": [],\n    \"section_301_applies\": false,\n    \"source\": \"rosetta_core\"\n  }],\n  \"entry_recommendation\": \"11\",  // Informal entry (<$2500)\n  \"isf_required\": false,  // Only for ocean\n  \"processing_time_ms\": 850\n}\n```\n\n### 3. Volume Estimates\n\nDHL Express handles ~5M US import shipments/year. At C8 integration:\n- **Phase 1 (pilot)**: 1,000 classifications/day (~30K/month)\n- **Phase 2 (rollout)**: 10,000/day (~300K/month)  \n- **Phase 3 (full)**: 50,000+/day (~1.5M/month)\n\nCurrent AskRosetta capacity: ~100 req/s on Cloud Run (auto-scaling).\n\n### 4. Pricing Model\n\n| Tier | Volume/month | Price/classification | Monthly |\n|------|-------------|---------------------|---------|\n| Pilot | Up to 30K | $0.15 | $4,500 |\n| Growth | 30K-300K | $0.08 | $8,000-$24,000 |\n| Enterprise | 300K+ | $0.03 | $9,000+ |\n\n**Value proposition for DHL**:\n- Current manual HTS classification: ~$2-5 per product (broker time)\n- C8 automated: $0.03-0.15 per product\n- **90-98% cost reduction** on classification\n- Reduced CBP penalties from misclassification\n- ISF automation for ocean shipments\n- PGA flag detection prevents holds at port\n\n### 5. Integration Architecture\n\n```\nDHL Systems → API Gateway → AskRosetta Cloud Run (Zurich)\n                                    ↓\n                          Rosetta Core (deterministic)\n                                    ↓ (fallback)\n                          Gemini Flash (AI classification)\n                                    ↓\n                          Results + Duty Calc → DHL\n```\n\n**Authentication**: API key (X-Api-Key header) with per-key rate limits and usage metering.\nAlready implemented in stripe_billing.py.\n\n### 6. SLA Commitments\n\n- **Latency**: <2s for single classification, <60s for 100-product batch\n- **Availability**: 99.9% (Cloud Run auto-scaling, multi-region possible)\n- **Accuracy**: >90% HTS accuracy at 6-digit level (with confidence scores)\n- **Data residency**: EU (Zurich) primary, US failover available\n\n### 7. Next Steps\n\n1. Request DHL API docs for their shipment data format\n2. Build DHL-specific adapter mapping their SKU format to ClassificationRequest\n3. Set up dedicated API key with DHL-specific rate limits\n4. Pilot with 100 real DHL shipments to calibrate accuracy\n5. Weekly accuracy review during pilot (target: 92%+ at 6-digit)"
}