Portfolio/MEM Executive Challenge

AI / ML  ·  Supply Chain  ·  Analytics

AI Supply Chain Visibility

MEM Executive Challenge, NC State  ·  Aug 2024 to Dec 2024

Role

Product Manager & Business Analyst

Type

Academic Competition

Client

AutoTech Motors (case study)

Outcome

Competed, did not place

View full documentation

The Challenge

A real business problem, treated like a real engagement

The MEM Executive Challenge is a competition where teams are given a complex enterprise case study and asked to develop a strategic recommendation. Our team was assigned AutoTech Motors, a fictional automotive manufacturer dealing with serious supply chain inefficiencies after COVID-era disruptions exposed how brittle manual forecasting really is.

The ask: design an AI-powered supply chain transformation strategy, including solution architecture, implementation roadmap, and projected ROI. I led the business analysis, solution framing, and final presentation.

The Problem

Manual forecasting, no real-time visibility, reactive inventory

Forecast method

Manual

spreadsheet-based, no ML

Supplier visibility

Reactive

delays found after the fact

Inventory approach

Historical

stockouts during demand spikes

AutoTech was losing revenue to stockouts during holiday demand spikes while simultaneously tying up capital in excess inventory during slow periods. The root cause: no predictive capability and no real-time supplier visibility. Every decision was made looking backward.

Our Recommendation

Four AI systems working together

Rather than proposing a single silver-bullet tool, we recommended a layered architecture connecting demand forecasting, supplier visibility, data integration, and inventory optimization.

01

Random Forest Demand Forecasting

Demand

ML model trained on historical sales, seasonality, and market data to predict demand with 80-85% accuracy, replacing manual spreadsheet forecasting.

02

SAP Real-Time Supply Chain Visibility

Visibility

SAP integration providing end-to-end visibility across global supplier networks with proactive disruption alerts instead of reactive scrambling.

03

Multi-Source Data Pipeline

Infrastructure

Continuous aggregation from suppliers, logistics partners, and market feeds feeding all AI models with clean, validated real-time data.

04

AI-Driven Inventory Optimization

Inventory

Dynamic safety stock and reorder point calculations balancing carrying costs against service level requirements, reducing excess without creating stockouts.

Projected Impact

The numbers we presented to the judges

80-85%

forecast accuracy

20-30%

inventory reduction

$2-3M

projected annual savings

50%

faster disruption response

Implementation Roadmap

A phased 18-month rollout

Months 1-3

Foundation

Form AI team, select vendors, set up infrastructure, run data quality audit.

Months 4-8

Integration

SAP integration, ML model training, workforce training, pilot with select product lines.

Months 9-12

Security & Governance

Cybersecurity framework, risk management, monitoring dashboards.

Months 13-18

Full Rollout

System-wide deployment, advanced analytics, continuous model improvement.

Reflection

What I took away from it

We did not place in the competition. Looking back, the strategy was thorough but the presentation lacked the sharpness and decisiveness that a panel of executives would respond to. We over-indexed on comprehensiveness and under-indexed on a clear, confident recommendation.

“Executives do not want every option on the table. They want you to tell them what to do and why. That is the lesson.”

Want to talk product?