FendrAI: Building an AI-Driven Risk Assurance Engine with Insights
Enterprise organizations are increasingly facing significant financial risks caused by hidden vendor overspend, billing errors, and non-compliance with contractual agreements. Traditional enterprise resource planning (ERP) systems often lack the intelligence and automation necessary to detect such discrepancies in real time. As a result, many organizations approve payments without properly validating invoice accuracy, leading to revenue leakage, prolonged disputes with vendors, and operational inefficiencies. This research project seeks to address these challenges by developing an AI-native model(s) designed to detect invoice anomalies and mitigate vendor overspending before payment approvals are processed. The proposed models integrate advanced natural language processing (NLP) techniques, machine learning (ML) models, and document intelligence technologies to build a robust solution capable of parsing complex invoices, analyzing contractual agreements, and detecting potential anomalies in real time. By doing so, the project aims to advance the state of the art in financial automation and risk management for large enterprises.
Voir la description complète du projetRasha Kashef
Fendr AI
Computer science
Information and cultural industries
Toronto Metropolitan University
Accelerate
