AI-Driven Insights from Knowledge Graphs via Integration with Large Language Models

In the current digital world, users are constantly sending companies signals of their experiences and preferences in various shapes and forms, including likes/dislikes of a particular offer/product, reviews on platforms and social media, and contacting customer service. However, despite the abundance of signals, E-commerce companies are unable to rapidly gain meaningful and actionable insights from those signals to react to the customer experience in a timely manner.

This project aims to address this challenge by leveraging knowledge graphs and Large Language Models (LLMs) to build an AI-based insights engine with a feedback loop to address the challenge of leveraging centralized customer signals/data from multiple sources to generate actionable insights that inform various product initiatives, campaigns and experiments across the organization to improve customer experience and generate business value.

The overarching goal is to provide a tool with an interface that allows business users to set their objectives or priorities and either automatically or manually prompt for hypothesis recommendations related to their product or campaign and receive insights for a complete set of hypotheses while accounting for overlapping initiatives within the company, enabling rapid experimentation.

Faculty Supervisor:

Ali Miri

Student:

Partner:

Rakuten Kobo Inc.

Discipline:

Computer science

Sector:

Manufacturing; Professional, scientific and technical services; Retail trade

University:

Toronto Metropolitan University

Program: