Supply chains generate enormous amounts of data — sales history, supplier performance, shipping records, market prices, geopolitical events — but most organizations analyze only a fraction of what’s available. AI enables supply chain teams to act on patterns that humans can’t consistently track.

1. Demand Forecasting

Statistical Demand Forecasting with AI

Prompt: Help me build a demand forecasting framework for our SKUs.

Business context:
- Industry: Consumer electronics retail
- SKUs: 2,400 active products
- Sales history: 36 months of daily data
- Seasonality: Strong Q4 (holiday), minor summer peak
- External factors: New product launches affect existing SKU demand
- Current method: Simple moving average (known to be inaccurate)

Help me:
1. Identify which forecasting models are appropriate for our data
   (ARIMA, Prophet, XGBoost, LSTM — and when to use each)
2. Design a segmentation approach (A/B/C items don't need same method)
3. List the external data sources that would improve accuracy
4. Define the right forecast horizon and granularity for our business
5. Establish error metrics (MAPE, RMSE, bias) and acceptable thresholds
6. Create a model selection framework for different SKU behaviors
7. Identify when AI forecasting outperforms simple methods vs. when it doesn't

New Product Forecasting

Prompt: Help me build a forecast for a new product launch.

Product: Premium wireless earbuds ($199 retail)
Launch: Q3 2026
Company: Mid-size consumer electronics company
Distribution: Major retailers + direct-to-consumer

Available data:
- Comparable product: Our previous earbud launch (SKU ABC, 2023)
  - Month 1: 8,200 units, Month 2: 5,400, Month 3: 4,100 (then stable at ~3,800)
- Market data: Category growing 18% YoY
- Competitor benchmark: Similar competitor product sold ~12K units/month
- Pre-orders: 3,200 units captured in 6-week pre-sale period
- Marketing budget: 30% higher than ABC launch

Build:
1. Base case monthly forecast (months 1-12)
2. Upside and downside scenarios with assumptions
3. Initial inventory recommendation for launch
4. Reorder trigger logic
5. Risk factors that could significantly change the forecast
6. What to monitor in months 1-2 to calibrate the forecast quickly

2. Supplier Risk Management

Supplier Risk Assessment

Prompt: Create a supplier risk framework for quarterly review.

Our situation:
- Total suppliers: 340 (direct materials)
- Critical suppliers (single-source or >15% of category spend): 28
- Geography: 60% Asia Pacific, 25% Europe, 15% Americas

Risk categories to assess each supplier:

Financial health:
- Credit rating trend
- Payment terms vs. actuals
- Recent earnings/news signals

Operational resilience:
- Geographic concentration of their facilities
- Backup manufacturing capacity
- Inventory they carry (days of supply)
- Lead time trend (improving/deteriorating)

Geopolitical and trade risk:
- Country risk score
- Tariff exposure
- Sanctions risk
- Regulatory compliance history

Quality and compliance:
- Defect rate trend
- Audit results (last 2 years)
- Corrective action closure rate
- Certification status

Concentration risk:
- % of our spend at this supplier
- Their dependence on us as customer
- Alternatives available (qualified backup suppliers)

Output: Risk score matrix and prioritized action list for top 10 risk suppliers

Disruption Scenario Planning

Prompt: Help me plan for these supply chain disruption scenarios.

Scenario 1: Port strikes at major West Coast ports (2-4 week duration)
Scenario 2: Key component supplier declares bankruptcy (Tier 1)
Scenario 3: Extreme weather event closes our primary DC for 3 weeks
Scenario 4: Regulatory change bans import of materials from Country X

For each scenario:
1. Immediate actions (0-72 hours)
2. Short-term response (1-4 weeks)
3. Medium-term recovery (1-6 months)
4. Long-term structural changes to prevent recurrence

For each action:
- Who is responsible?
- What decisions need executive approval?
- What triggers this escalation?
- What data do we need to monitor?

Also: Pre-event investments that reduce impact of each scenario

3. Inventory Optimization

Safety Stock Calculation

Prompt: Help me calculate safety stock for our inventory system.

Context:
- We carry 800 SKUs across 3 distribution centers
- Current approach: Static "4 weeks of supply" safety stock for everything
- Known problem: Stockouts on fast movers, overstock on slow movers
- Fill rate target: 99% for A items, 97% for B items, 95% for C items

For a proper safety stock model, help me:

1. Define the statistical safety stock formula appropriate for our situation
   SS = Z × σ_LTD where Z is service level factor, σ_LTD is demand 
   variability over lead time

2. Build the calculation framework:
   - How to calculate demand standard deviation from our history
   - How to incorporate lead time variability (not just demand variability)
   - How to adjust for item classification (A/B/C)

3. Example calculation:
   SKU: Widget Pro
   Average daily demand: 45 units
   Demand standard deviation: 12 units/day
   Average lead time: 14 days
   Lead time standard deviation: 2 days
   Target service level: 99% (Z = 2.33)

4. When to recalculate (event triggers vs. calendar schedule)
5. How to handle seasonal items (different SS for different seasons)

Excess Inventory Liquidation Planning

Prompt: Help me create a liquidation plan for excess inventory.

Situation:
- Total excess inventory: $4.2M at cost
- Categories:
  - Category A ($1.8M): Slow-moving but still selling
  - Category B ($1.4M): Not selling for 6+ months
  - Category C ($1.0M): Obsolete (product discontinued)
  - Seasonal ($0.4M): Off-season inventory

Constraints:
- Can't disrupt primary channel (retail) pricing
- Warehouse space is costing $85K/month to hold this
- Accounting wants to clear before fiscal year end (4 months)

Options to evaluate:
1. Markdowns through primary channel (timing and depth)
2. Secondary marketplaces (overstock.com, Amazon warehouse deals)
3. Off-price channel partnerships
4. B2B bulk sale to distributors
5. Donation (tax benefit vs. cash recovery)
6. Disposal (cost vs. recovery)

For each option:
- Expected recovery % of cost
- Timeline to clear
- Channel conflict risk
- Administrative burden
- Recommendation by category

Create: Prioritized liquidation plan with 30/60/90-day targets

4. Logistics and Transportation

Carrier Selection and Routing

Prompt: Help me optimize our carrier mix and routing strategy.

Our shipping profile:
- Volume: 12,000 shipments/month
- Geography: Nationwide US (ship from 2 DCs: Ohio and California)
- Package mix: 
  - 40% packages under 5 lbs (express eligible)
  - 35% packages 5-20 lbs (ground)
  - 25% packages over 20 lbs (freight)
- Customer requirements: 2-day delivery for 80% of orders

Current carrier mix:
- UPS: 65% of volume, negotiated rate
- FedEx: 25% of volume
- USPS: 10% (last mile for rural)
- No regional carriers currently

Analyze:
1. Which orders should route to which carrier based on zone + weight?
2. Where do regional carriers (OnTrac, Laser Ship, LSO) outcompete UPS/FedEx?
3. What additional volume threshold unlocks better negotiated rates?
4. Potential savings from zone-skipping (DC 2 ships orders destined for zones 1-3 from nearby)
5. Returns network optimization (are we paying too much on returns?)

Output: Recommended carrier allocation by zone and weight band with savings estimate

Freight Audit Framework

Prompt: Build a freight invoice audit process.

Current situation:
- Annual freight spend: $2.8M
- Invoices: 4,500/month across 6 carriers
- Current audit: Manual spot check of 5%
- Known error types: 
  - Address correction fees (many disputes)
  - Dimensional weight miscalculation
  - Wrong rate class applied
  - Duplicate billing
  - Missing contract discounts

Design:
1. Automated audit rules (what % should be flagged automatically?)
2. Which error types are highest ROI to audit?
3. Claim process for each carrier
4. Staff time vs. expected recovery analysis
5. Tools recommendation (freight audit software vs. in-house)
6. Expected recovery rate benchmarks
7. Contract negotiation leverage from systematic audit

5. Procurement AI Applications

RFQ and Supplier Negotiation

Prompt: Help me prepare for a sourcing negotiation with our key packaging supplier.

Context:
- Current contract: $2.1M/year, expires in 3 months
- Supplier: Single-source for 3 specialty packaging items
- Market: Packaging costs down 8% industry-wide this year
- Our leverage: 12% volume increase we can offer; competing bid in hand
- Their leverage: 14-week tooling lead time; relationship dependencies

Prepare:
1. Opening position and anchoring strategy
2. Target price reduction (what's realistic given market data?)
3. Non-price terms to negotiate (payment, MOQ, lead time, quality guarantees)
4. BATNA analysis (what's our best alternative if they don't move?)
5. Concession sequence (what do we give up, in what order?)
6. Red lines (what would make us walk away?)
7. Language for the opening: Set the tone without being adversarial

Negotiation framework: Interest-based negotiation (find joint value, not just split the pie)

Spend Analysis

Prompt: Analyze this AP spend data and identify savings opportunities.

Data: [Paste 12 months of PO data: vendor, category, amount, frequency]

Analyze for:
1. Spend concentration
   - Top 20 vendors by spend
   - % of spend with top 10 suppliers
   - Long tail vendor count and spend

2. Category analysis
   - What we spend in each category
   - Fragmentation within categories (too many suppliers?)
   - Categories ripe for consolidation

3. Maverick spend
   - Purchases outside preferred suppliers
   - Purchases outside contract pricing
   - P-card spend vs. PO (where are the controls gaps?)

4. Price variance
   - Same item, different prices (price leakage)
   - Off-contract purchasing patterns

5. Savings opportunities ranked by:
   - Consolidation opportunities ($X by consolidating Y vendors)
   - Negotiation leverage (volume + fragmentation)
   - Quick wins vs. strategic initiatives

Output: Savings opportunity register with estimated value, timeline, and owner

AI in supply chain is most powerful when it surfaces patterns in large, multi-dimensional datasets that human analysts can’t consistently process — the final judgment on supplier relationships, risk tolerance, and strategic trade-offs remains with experienced operations leaders.