Skip to main content

DEMO PREPARATION GUIDE - ChainAlign Decision Intelligence Platform

Purpose: Prepare ChainAlign for live demonstration to stakeholders Target Audience: Sales & Operations Planning (S&OP) practitioners Demo Duration: 30-45 minutes Key Differentiation: Explains demand movements through external factors (weather, policy, economics)

Pre-Demo Setup Checklist​

Backend Services​

  • Database migrations run (npx knex migrate:latest)
  • Environment variables set (.env configured)
    • FIREBASE credentials
    • GEMINI_API_KEY (for AI explanations)
    • Database credentials
    • Optional: FRED_API_KEY (for US economic data)
  • Backend server running (npm start)
  • All API routes responding (test /api/test-route)

Frontend Setup​

  • React dev server running (npm start from frontend/)
  • Build without errors
  • Auth context configured
  • API base URL pointing to backend

Demo Data Preparation​

  • Create test tenant and user
  • Load sample demand history (CSV with dates and demand values)
  • Load sample weather data (or sync from Open-Meteo)
  • Create sample policy events
  • Create sample annotations and comments

Network & Access​

  • Ports accessible (8080 for backend, 3001 for frontend)
  • No firewalls blocking APIs
  • Backup internet connection available
  • Demo URL bookmarked and tested

Demo Flow (30-45 minutes)​

1. Opening - Problem Statement (2 min)​

Narrative: "Traditional S&OP is reactive. You wait for demand to change, then react."

Demo:

  • Show dashboard with historical demand data
  • Point out: "Here's 2 years of daily demand"
  • Question: "But WHY did demand spike on March 15th? Why the drop in July?"

Transition: "Most companies don't know. But we can discover it."

2. Weather Correlation Discovery (6 min)​

Narrative: "What if we discovered that temperature correlates with demand?"

Demo Steps:

  1. Show Weather Historical Data

    GET /api/weather/historical/40.7128/-74.0060
    ?start_date=2019-01-01&end_date=2024-10-22
    • Display 5 years of daily temperature data
    • Show quality scores (0-1, reliability metric)
    • Point out: "This is real historical data from Open-Meteo API"

  2. Trigger Correlation Discovery (note: this would be manual in demo)

    • Input: 5 years of demand history + 5 years of weather
    • Show process: "Analyzing correlations across all lags (0-7 days)"
    • Show results: 
      {
        "weather_metric": "temp_avg",
        "correlation_coefficient": 0.72,
        "p_value": 0.0001,
        "lag_days": 0,
        "r_squared": 0.52,
        "equation": "demand = 3.2 × temp + 1000",
        "confidence": "95% (p < 0.01)"
      }

  3. Explain Finding

    • "72% correlation: strong relationship"
    • "P-value < 0.001: 99.9% confidence this isn't random"
    • "Explains 52% of demand variance"
    • "Every 1°C increase: expect +3.2% demand"

  4. Show Feature Importance Wow Card

    • Display the FeatureImportanceWowCard component, highlighting weather_temp as a top impactful feature.
    • Refined Explanation: "After the system analyzes the correlations, it doesn't just give you a number. It shows you why it believes certain factors are important. This 'Wow Card' visually summarizes the most impactful features, like weather temperature, and quantifies their contribution to demand. It's our way of making the AI's insights transparent and actionable."

  5. Show Forecast Application

    POST /api/weather/forecast
    {
      "location_id": "NYC",
      "weather_forecast": [
        {"date": "2024-10-23", "temp_avg": 18},
        {"date": "2024-10-24", "temp_avg": 22}
      ],
      "demand_baseline": 1000
    }
    • Response: 
      {
        "date": "2024-10-24",
        "adjusted_demand": 1154,
        "adjustment": "+15.4%",
        "confidence_interval": ["+10.2%", "+20.6%"],
        "explanation": "Based on 5-year history, temp rise correlates with demand"
      }

Key Takeaway: "System found a relationship that explains demand movements."

3. Data Quality Gates (5 min)​

Narrative: "Ensuring high-quality data is paramount. Our robust data infrastructure and quality gates are designed to prevent issues, even when underlying systems face challenges, guaranteeing reliable inputs for your forecasts."

Demo Steps:

  1. Show Annotations System

    • Display sample annotation: "HIGH severity: Demand data incomplete for March 1-5"
    • Show severity levels: HIGH (blocks), MEDIUM (warns), LOW (info)

  2. Show Data Freshness Indicators

    GET /api/freshness/summary
    • Response: "72% FRESH, 18% RECENT, 8% STALE, 2% CRITICAL"
    • Show visual: Green (fresh), blue (recent), orange (stale), red (critical)

  3. Show Comments for Collaboration

    • Public comment: "We had supply chain disruption that week"
    • Personal note: "Need to investigate with IT"
    • Show edit history and timestamps

  4. Show Confirmation Workflow

    GET /api/confirmation/readiness/DEMAND_DATA
    • Shows readiness: "BLOCKED" if HIGH annotations or critical data freshness
    • Lists blocking issues with details
    • Option to override with documented reason
    • Audit trail: shows who approved and when

Key Takeaway: "Quality gates prevent bad forecasts. Collaboration prevents surprises."

4. Policy & Economic Intelligence (8 min)​

Narrative: "Supply chains don't react to demand alone. They react to policy changes."

Demo Steps:

  1. Create Policy Event

    POST /api/policy/events
    {
      "event_type": "TARIFF_CHANGE",
      "event_date": "2024-03-15",
      "announced_date": "2024-01-15",
      "description": "25% tariff on imports from China",
      "affected_countries": ["CN"],
      "affected_hs_codes": ["290511"],
      "impact_magnitude": "25%"
    }
    • Created event shows up in timeline

  2. Show Economic Indicators

    GET /api/economic/imf
    ?countries=US,CN&indicators=NGDP_RPCH,PCPIPCH
    • China GDP growth: 7% → 5.2% (declining)
    • "Economic slowdown explains lower Chinese supplier capacity"

  3. Track Routing Change

    POST /api/policy/routing
    {
      "supplier_id": "ABC-123",
      "old_destination": "CN",
      "new_destination": "VN",
      "policy_event_id": "EVT-123",
      "cost_impact": -8.3,
      "lead_time_impact": 7,
      "volume": 10000
    }
    • Routing change linked to policy event
    • Cost and lead time impacts recorded

  4. Analyze Policy Impact

    GET /api/policy/impact/EVT-123
    • 45 SKUs rerouted
    • Total volume: 120,000 units/year
    • Average cost impact: +8.3%
    • Average lead time: +5 days

  5. Show Policy Timeline

    GET /api/policy/timeline?months=24
    • Chronological view of policy events
    • Lead time from announcement to routing change
    • Impact analysis per event

  6. Show Response Lead Time Analytics

    GET /api/policy/response-time
    • "Supply chain responds in ~45 days average"
    • Breakdown: TARIFF_CHANGE (45d), FTA (32d), SANCTIONS (62d)

Key Takeaway: "Policy changes drive routing decisions. We track and measure the impact."

5. Integrated Forecast with External Factors (5 min)​

Narrative: "Now we combine everything: weather + policy + economics."

Demo Steps:

  1. Show Comprehensive Forecast

    • Temperature forecast: +5°C next week
    • Policy event: New tariff announced
    • Economic indicator: GDP declining
    • Reasoning section shows all factors

  2. Show Feature Importance Wow Card (again)

    • Re-display the FeatureImportanceWowCard component, emphasizing how the system uses these factors to adjust the forecast.
    • Refined Explanation: "Notice how the 'Wow Card' now reflects the influence of multiple external factors. This isn't just a black box; ChainAlign shows you the precise drivers behind its forecast adjustments, giving you confidence in the recommendations."

  3. Explain Forecast Adjustments

    Demand baseline: 1,000 units/week
    Weather adjustment: +15% (temperature correlation)
    Policy adjustment: -5% (tariff concerns, higher prices)
    Economic adjustment: -3% (declining GDP)
    Final forecast: 1,070 units/week (+7%)
    Confidence: 73% (based on 5-year history)

  4. Show Confidence Factors

    • Correlation count: 3 (good)
    • Correlation strength: 0.65 average (strong)
    • Adjustment agreement: 0.82 (consistent)
    • Overall confidence: MEDIUM-HIGH

Key Takeaway: "System explains forecast through multiple external factors."

6. Real-World Scenario: Pharma API Supply Chain (5 min)​

Your Example - Use this to connect with pharma industry audience

Scenario: "Pharma API manufacturer supplying US market from China"

Demo Setup:

  • SKU: API-ACETAMINOPHEN
  • Annual volume: 500 tons
  • Current sourcing: 100% China (20% tariff)
  • Alternative sources: India (12% cost premium), Vietnam (15% premium)

What-If Scenario:

Policy Event: US tariff on Chinese APIs increases from 20% → 25%
System predicts:
  - Cost impact: +$1.2M annually (additional tariff)
  - Alternative cost: India rerouting costs $1.8M
  - Net savings: Use India supply (+$600K benefit annually)
  - Lead time: 45 days to reroute (affects short-term contracts)

Recommendation: "Secure India supplier now, reroute supply in 45 days"

Show in Dashboard:

  • Current supply chain routing
  • Cost comparison table
  • Timeline to implementation
  • Risk factors (supplier capacity, quality)

Key Takeaway: "System guides S&OP decisions through data-driven insights."

7. Closing - Differentiation (2 min)​

Narrative: "Why ChainAlign is different"

Key Points:

  1. Discovers vs Assumes: Finds actual correlations from your data
  2. Self-Optimizing Roadmap: Automatically identifies and prioritizes impactful features based on their predictive lift, even generating high-priority Linear issues for your team when critical insights are found.
  3. Explains vs Black Box: Shows reasoning: "Because X happened, Y will happen"
  4. Integrates: Combines weather + policy + economics + internal demand data
  5. Transparent: Full audit trail of decisions, annotations, comments
  6. Real-Time: Live dashboards, automatic data freshness tracking
  7. Extensible: Can add new external data sources (news, social media, etc.)

Call to Action:

  • "Let's start with your pilot data"
  • "What SKUs would benefit most from this analysis?"
  • "We can discover correlations specific to your business"

Demo Data Files​

Sample Demand History (CSV)​

date,demand
2019-01-01,950
2019-01-02,1020
2019-01-03,1050
...
2024-10-22,1150

Sample Annotations (Pre-created in DB)​

{
  "data_element_type": "DEMAND_DATA",
  "severity": "HIGH",
  "description": "Data incomplete for March 1-5 due to system outage",
  "created_by_name": "Alice Johnson",
  "created_at": "2024-10-20"
}

Sample Policy Events (Pre-created in DB)​

{
  "event_type": "TARIFF_CHANGE",
  "event_date": "2024-03-15",
  "announced_date": "2024-01-15",
  "description": "US tariff on Chinese chemicals increased 25%",
  "affected_countries": ["CN"],
  "impact_magnitude": "25%"
}

Sample Comments (Pre-created in DB)​

{
  "commentable_type": "ANNOTATION",
  "content": "We had a supply chain disruption that week due to port congestion",
  "is_public": true,
  "created_by_name": "Bob Smith"
}

Common Questions & Answers​

Q: How accurate are the weather correlations? A: Accuracy depends on your data quality. Our demo shows r=0.72 correlation (52% variance explained). Your data might show different patterns. That's the value - we discover YOUR correlations.

Q: What if we don't have 5 years of data? A: Minimum recommended is 2 years (730 data points). Fewer data points = lower confidence. System shows confidence intervals based on sample size.

Q: Can you integrate other external data? A: Yes. Weather (done), policy (done), economics (done). We can add: news sentiment, social media, competitor actions, seasonal patterns, custom indicators.

Q: How do you handle seasonality? A: We can analyze by season (summer vs winter) or build separate profiles. Demo focuses on overall correlation, but granular analysis is possible.

Q: What about lag effects? A: We test 0-7 day lags. For example: "Tariff announced Monday, routing change decision made Friday (4-day lag)."

Q: Is this replacing our forecast model? A: No. It's enhancing it. Your existing forecast + external factor adjustments = better forecast. System transparency shows the adjustments.

Troubleshooting During Demo​

API Returns 401 Unauthorized​

  • Check: Is token being passed in request?
  • Check: Is user logged in?
  • Fix: Log out and log back in

Weather Data Returns Empty​

  • Check: Date range is valid (YYYY-MM-DD format)
  • Check: Latitude/longitude are valid (e.g., 40.7128/-74.0060 for NYC)
  • Fix: Start with pre-cached NYC data: 40.7128/-74.0060

Forecast Confidence is LOW​

  • Reason: Few correlations found, or correlations are weak
  • Explain: "This shows the system is conservative - only forecasts when data supports it"
  • Workaround: Show pre-calculated demo forecast instead

Policy Timeline Returns Empty​

  • Reason: No policy events created yet in demo tenant
  • Fix: Pre-create policy events before demo starts
  • Workaround: Use provided sample policy event

Charts Not Loading​

  • Check: Frontend is connected to correct backend URL
  • Check: Browser console for errors
  • Fix: Restart frontend dev server

Success Metrics​

Demo is successful if:

  • ✅ Audience understands weather correlation discovery
  • ✅ Audience sees value in policy event tracking
  • ✅ Audience asks: "Can you do this for our products?"
  • ✅ Audience understands explainable AI (system shows reasoning)
  • ✅ Audience asks about pilot/POC

Red flags during demo:

  • ❌ API errors (401, 500) - technical credibility issue
  • ❌ Confusing explanations - need simpler narrative
  • ❌ Slow performance - optimization needed
  • ❌ Missing data - pre-load demo data properly

Post-Demo Follow-Up​

Immediate (Day 1)​

  • Send demo recording/slides
  • Provide pricing & engagement model
  • Outline pilot program (3-6 months)

Within 1 Week​

  • Pilot proposal: Which SKUs/products?
  • Data requirements: Demand history, supply chain, policy events
  • Timeline: Implementation and first results

Pilot Phase (3 months)​

  • Load customer's real demand data
  • Discover actual correlations for their products
  • Build forecasts with external factor adjustments
  • Compare to their current forecast model
  • Measure accuracy improvements

Demo Success Checklist​

Pre-Demo:

  • Backend running and tested
  • Frontend running and responsive
  • Demo data loaded
  • All APIs responding
  • Network connectivity verified
  • Backup demo recording available
  • Slides prepared
  • Q&A talking points documented

During Demo:

  • Start with problem statement
  • Progress through demo flow sequentially
  • Pause for questions (don't rush)
  • Use pharma example to connect with audience
  • Show confidence intervals (not just point estimates)
  • Highlight transparency (audit trail, explanations)
  • Demonstrate the "Frontend Wow Card" to visualize feature importance insights

Post-Demo:

  • Capture contact info
  • Send follow-up within 24 hours
  • Address any technical questions
  • Propose next steps (pilot program)

Ready to demo! 🚀

Key message: "ChainAlign explains demand movements through external factors your competitors don't measure."