Build a GenAI-Based Procurement Tool Featuring Natural Language Queries
Introduction
Imagine real-time data feeds from an ERP system accessible through AI-assisted queries and responses. What could that do? It could help give procurement management teams a better understanding of what’s happening with purchase orders sent to a supplier in near real time. Likewise, it could provide an easy-to-use UI for suppliers to know the status of payable invoices raised to a buyer. It could also be applied to other ERP documents such as payable invoices, receivables, and so on. It’s adaptable to recruitment, customer experience apps, and more. It could even be used on other customer query channels, including social media platforms and web apps.
This project explores how to make a tool that helps enable real-time purchase order feeds (updates, creations, and so forth) coming from an ERP system (Oracle Cloud ERP in this case) into a procurement database store, transform procurement queries in natural language to SQL queries, and synthesize the SQL responses using Oracle Cloud Infrastructure (OCI) Generative AI, Oracle Integration Cloud, and Oracle Autonomous Database.
Demo
Prerequisites and setup
- Oracle Cloud account—sign-up page
- Getting started with OCI Generative AI—documentation
- OCI Generative AI—Python SDK