AI IN SUPPLY CHAIN MANAGEMENT

Meet Bradley Rogers, walk through the AI Supply Chain Implementation Portfolio, and get oriented on assignments and expectations before the course begins.

• Instructor introduction
• Course and assignment overview
• Final deliverable overview: AI Implementation Portfolio

Explore how supply chain AI evolved from predictive models to autonomous agents, and see where organizations are capturing real ROI versus stalling at the pilot stage.

• History of automation and AI in supply chain
• Predictive ML versus agentic AI
• Where AI delivers ROI today
• Why some AI initiatives stall
• Course structure and deliverables
• Case study: Amazon's AI-driven fulfillment network at 520,000 robots and 400 million daily demand predictions

Learn what large language models and Retrieval-Augmented Generation actually do, and identify which type of AI agent fits which supply chain task.

• Large language models without the math
• Retrieval-Augmented Generation basics
• Reactive, tool-use, and multi-agent types
• Matching AI type to workflow
• Demo: First look at the Chainlytix AI Readiness Diagnostic portal

Use the SCOR framework to sort AI opportunities into green, yellow, and red zones, and learn the constraints that stall AI before it scales.

• SCOR framework for organizing AI opportunities
• Green, yellow, and red readiness zones
• Six constraints that stall AI
• Free synthetic data via Kaggle
• Demo: Scoring a real supply chain task with Chainlytix as a class
• Case study: MIT NANDA's finding that 95% of enterprise AI pilots show no P&L impact

Assignment #1: The AI Opportunity Matrix
Identify three supply chain bottlenecks and propose a high-level AI solution for each, backed by a clear business case.

Learn what AI agents need from your data versus what dashboards need, and run a hands-on exploratory data analysis on your own supply chain data.

• What AI agents need from data
• Data quality, availability, and integration blockers
• Preprocessing techniques for supply chain data
• Exploratory data analysis fundamentals
• Workshop: Evaluating your own data for AI readiness using Chainlytix

Design governance frameworks that set clear boundaries for AI autonomy, and see how enterprise platforms like IBM Watsonx put responsible AI governance into practice.

• Governance as a business enabler
• Ethical boundaries and human oversight
• Data privacy and compliance considerations
• IBM Watsonx governance example
• Case study: Knight Capital's $440M loss in 45 minutes from a deployment with no kill switch
• Guest speaker: Senior enterprise AI governance practitioner

Assignment #2: Data Readiness & Workforce Impact Audit
Audit the data each of your three AI use cases would need, and assess how implementation would shift your team's day-to-day work.

Build predictive demand forecasting models with AI platforms, and learn why even accurate forecasts sometimes still fail to drive better inventory decisions.

• AI demand forecasting fundamentals
• Managing demand variability
• Integrating forecasts with planning software
• Measuring forecast accuracy and ROI
• Demo: KNIME walkthrough for AI-driven demand forecasting
• Case study: Zillow Offers' forecasting model and its data-drift failure

Use demand forecasting insights to optimize inventory levels, and learn when to trust automated replenishment and when to override it safely.

• AI-driven inventory management
• Reducing carrying costs with forecasting insights
• Automated replenishment triggers and overrides
• Integrating AI with WMS and ERP
• Case study: JD.com's end-to-end deep learning inventory model

Assignment #3: The Pilot Selection
Evaluate your three use cases and select the one with the strongest chance of success, based on operational readiness and ease of adoption.

Explore why route optimization is one of AI's most mature autonomous use cases, and see how IoT sensors power real-time supply chain visibility.

• Matrix and isoline routing techniques
• IoT sensors as the visibility backbone
• AI tracking and tracing integration with ERP
• Scaling across multi-region operations
• Case study: Ocado's warehouse robot fires and the limits of scaling without safety architecture
• Guest speaker: Real-time visibility platform such as Project44 or FourKites

Automate order tracking and escalation workflows with natural language processing, and learn where chatbots make sense internally versus in customer-facing supply chain roles.

• NLP-based customer service applications
• Internal versus customer-facing chatbot design
• Automating order tracking and escalation
• Enriching responses with Tavily's real-time data
• Case study: McDonald's and IBM's voice-ordering rollback
• Case study: Klarna's AI customer service, from the win to the walk-back

Plan, build, and stress-test your own supply chain chatbot using prompt engineering, guardrails, and approval workflows for safe deployment.

• Scoping your chatbot's use case
• Prompt engineering and knowledge boundaries
• Guardrails and approval workflows
• Workshop: Building an internal Supply Chain Knowledge Assistant with Make.com

Assignment #4: Build & Validate an Internal Supply Chain Knowledge Assistant
Design and build a no-code chatbot for your pilot, complete with a knowledge base, guardrails, and realistic test scenarios.

Apply predictive analytics to identify and prioritize supply chain risks, and design human-in-the-loop workflows for decisions that still need judgment.

• Predictive analytics for risk prioritization
• Human-in-the-loop decision design
• AI-powered disruption monitoring and supplier intelligence
• Demo: Tavily walkthrough for live web monitoring
• Case study: Zillow Offers' model risk and its $500M write-down
• Guest speaker: Tavily's Laney and one of her engineers

Design and stress-test an autonomous AI agent that retrieves data and monitors risk, then connect it to your team's real communication channels.

• Agents versus chatbots: tools, memory, action
• Defining tasks, tools, and guardrails
• Connecting outputs to email, Slack, or dashboards
• Comparing agents with platforms like UiPath
• Workshop: Building a risk-monitoring agent with Make.com and Tavily

Assignment #5: Build & Test a Supply Chain AI Agent
Build and evaluate an autonomous agent that automates a real operational task, then develop a 90-day roadmap for deploying it responsibly.

Translate your AI outputs into a business case, and practice presenting your portfolio and prototype to the class for executive-style feedback.

• Translating outputs into investment decisions
• Demonstrating ROI and time-to-value
• Presenting to executives, IT, and frontline teams
• Presenting your portfolio and agent or chatbot to the class
• Guest speaker: Andrew Kerr, VP at AlixPartners

Position your new AI skills, portfolio, and agent build for the roles employers are hiring for now in supply chain AI.

• Emerging supply chain AI roles
• Positioning your portfolio on your resume
• Live final project showcase
• Emerging trends in agentic supply chain AI