AI Purchase Order Automation Software for Streamlined Processing and Enhanced Efficiency

Customer Overview

Our client is a US-based distributor, partnering with a leading global manufacturer of fluid system components. They provide a wide range of high-quality products, including fittings, valves, regulators, gauges, hoses, and tubing, to meet the needs of various industries. The distributor works closely with customers to offer tailored quotations and ensures timely delivery of products against their Purchase Orders.

Project Overview

The distributor receives many Inquiries and Purchase Orders from customers, each containing hundreds of products, daily. Their Customer Service and Sales teams use Zendesk and SAP to process these 10-to-15-page-long Inquiries and POs. The teams manually managed the workflow, leading to slow response times and frequent errors.

The key opportunity was to automate these tasks to reduce turnaround times, minimize mistakes, and enable faster order deliveries.

Challenges

Building an automation software to shorten the turn-around time for handling Inquiries, processing POs, and fulfilling orders while eliminating errors and improving efficiency.

• For each customer Inquiry or PO, the customer service team had to manually download the documents, create Zendesk tickets, read the entire PO, and create a Sales Order in SAP.
• As the workflow between the teams to handle Inquiries or POs was semi-automated, it caused communication gaps and reduced productivity.
• For all Inquiries or POs, the sales team had to manually gather product information and each product’s availability for fulfillment by referring to their different systems. It took several hours to collect this information.
• The customer service team manually entered lengthy PO details in SAP, leading to erroneous item quantities, price, delivery location & time, etc. The distributor wanted to shorten the time between PO receipt & SAP entry from 4 days to half an hour.
• The distributor wanted a solution that integrates multiple systems like email, chat, Zendesk, SAP, and database to automate processes.

Solution

We built an Inquiry & Purchase Order Automation Software using multiple AI Agents for document understanding, data extraction, and generating Zendesk tickets and Sales Orders in SAP.

• We developed an Agentic framework using LangChain Graph to deploy multiple AI agents, including a Coordinator Agent and task-specific agents, each handling distinct responsibilities to automate Inquiry and PO processing workflows.
• The Coordinator Agent manages the overall process, assigning tasks to specialized agents and ensuring the correct execution of each task in the proper sequence.
• The task-specific agents follow the Coordinator Agent's instructions to complete specialized tasks, such as classifying documents, summarizing key data, and integrating with systems for data processing.
• After completing their tasks, the agents report back to the Coordinator Agent for further processing or confirmation of completion.
• To ensure optimal AI performance and early issue detection, we implemented human-in-the-loop techniques for oversight and approval, ensuring nothing was executed without human verification.

Benefits

The Purchase Order Automation Software we developed, delivered the following benefits:

• Automation reduced inquiry response time by 20x and PO entry time by 98%, enabling staff to focus on strategic activities, and boosting profitability.
• The AI-based solution with human-in-loop techniques reduced errors by 70%, improving operational accuracy and minimizing costly error management.
• Integration of systems and vendor ordering improved operational efficiency by 20x, allowing the team to handle more orders and increase sales volume.
• Automating repetitive tasks like generating quotations and sales orders reduced the billing cycle by 85%, improving cash flow and payment collection delays.

Technology

• Python
• FastAPI
• AWS Bedrock
• AWS S3
• Claude 3.5
• LangChain Graph

Industry

• Manufacturing (Fluid System Components)