Building Retail Forecasting Agents using Copilot for Modern E-commerce Inventory Management

A practical exploration of hybrid AI-powered forecasting techniques tailored for e-commerce businesses facing unique supply chain challenges and demand patterns. This talk demonstrates how Copilot can be leveraged to build retail forecasting agents that account for distinct seasonalities, product cycles, and regional shopping behaviors. Learn how companies like Target and their vendor brands transform complex data into actionable inventory insights, balancing interpretability needed for effective cross-functional communication between data science and business teams in the modern marketplace powered by deep learning systems.

Data Science Salon 2025

Traditional Time Series in Modern Times – A Retail Forecasters Guide

This talk explores the evolution of time-series forecasting in retail, highlighting how traditional statistical approaches can be integrated with modern neural network techniques to create models that are both precise and transparent. It discusses the challenges of using complex, black-box models in environments where understanding the influence of seasonality, holiday effects, and pricing dynamics is essential for collaborative decision-making.