Skip to content

Procurement Partners Streamlines Supply Chain and Spend Management with Generative AI

Executive Summary

Procurement Partners needed a way for customers to automate the manual assignment of general ledger (GL) codes to purchased items. Mission Cloud developed a solution using generative AI from Amazon Bedrock and Anthropic Claude. By integrating with Procurement Partners' user chat interface, the solution analyzes and matches purchased items with the appropriate GL codes.

This automation improves operational efficiency, accuracy, and cost savings for Procurement Partners and its customers. The software uses Amazon Web Services (AWS) offerings such as AWS Fargate and Spot Instances to ensure cost efficiency and eliminate the complexity of server management.

About the Customer

Procurement Partners is on a mission to simplify procurement and payment processes by providing reliable, compliant, and transparent solutions that prioritize efficiency. With a focus on healthcare supply chain and spend management, the company offers procurement software to help companies optimize medical inventory management, purchasing and accounts payable tasks.

Background: Accelerating Procurement Processes

Healthcare companies require efficient inventory management, accurate purchasing, and timely payment processing. However, manual processes and outdated systems create inefficiencies, errors, and delays, affecting ‌healthcare providers and their patients.

Recognizing these challenges, Procurement Partners provides procurement teams with a unified platform that improves operational efficiency and reduces errors. The company sought to improve operational efficiency with generative AI and co-sponsored a hackathon with Mission Cloud to identify potential opportunities and use cases. 

Challenge: Automatically Classifying Items with AI

This collaborative effort revealed pain points, including the manual assignment of general ledger (GL) codes to purchased items. GL codes are unique identifiers for categorizing and tracking financial transactions, and they’re crucial for organizing and analyzing procurement data, in addition to tracking expenses against budget.

The existing manual approach required individually assigning GL codes to each purchased item, and each purchase order could contain over 100 line items. This manual approach drained resources and reduced operational efficiency. With a growing customer base and increasing transaction volumes, Procurement Partners recognized the need to automate this process. Doing so would help its customers streamline operations, reduce costs, and improve accuracy.

The risk of erroneous GL codes with this manual process could also negatively affect financial reporting and analysis, which would result in overspending. The challenge was to develop a solution that could automatically analyze and match purchased items with the appropriate GL codes, thus giving customers an optimized supply chain and spend management process. 

Solution: Using Amazon Bedrock and Claude for Analysis

Mission Cloud developed a solution that uses generative AI powered by Amazon Bedrock and Anthropic Claude to analyze and match purchased items with the appropriate GL codes. The model is trained on historical data and uses advanced algorithms to learn patterns and make accurate predictions for GL code assignment. The solution integrates with Procurement Partners' existing user chat interface and provides an API so customers can automate the GL code assignment process.

When customers upload an item, the solution's algorithms analyze data points, including information about the item, customer, and facility, and generate a ranking-based estimation of the appropriate GL code. The response from the API includes the item's category (e.g., medical, COVID, etc.) and the corresponding GL code.

The solution also employs a database and caching mechanism to improve the accuracy and efficiency of GL code assignment. The database stores all categories related to a customer and facility, enabling quick retrieval and matching. 

Results: Efficient, Accurate, and Cost-Effective Code Assignments

Procurement Partners has gained a market-ready solution that empowers its customers to automate GL code assignments. Manual GL code work is eliminated, allowing customers to focus on other areas of procurement and improving efficiency and productivity within the healthcare supply chain. More specifically, the solution significantly reduces the amount of time it takes to place a purchase order as well as the corresponding invoice that is received. Additionally, Procurement Partners customers have a more accurate picture of their budgets throughout the month, resulting in better budget management.

The solution also ensures consistency in GL code assignment, minimizes errors, improves accuracy, and creates cost savings for Procurement Partners and their customers. It was designed with cost efficiency in mind. For example, using AWS services such as AWS Fargate and Spot instances eliminates the complexity of provisioning and managing servers and helps Procurement Partners avoid unnecessary costs.

AWS Services Used

During the engagement with Mission Cloud, Procurement Partners used a variety of AWS and third-party services, including: 

Amazon Bedrock
Anthropic Claude
Amazon ECS

Elastic Load Balancing

Amazon S3


AWS Fargate