The University of Michigan’s Shared Services Center (SSC) is embracing AI to streamline operations and improve efficiency. Processing approximately 300,000 invoices annually, the SSC Accounts Payable team faced a significant challenge: extracting key information from vendor-submitted vouchers while relying on a template-based system that only automated about 30% of invoices. With this older technology being phased out, SSC saw an opportunity to modernize its workflow with an AI-driven solution designed to enhance accuracy, efficiency, and scalability.
The core of this transformation was realized by leveraging the Large Language Models available in the U-M GPT Toolkit. When a vendor submits an invoice, the AI extracts key details such as invoice number, vendor information, and payment amounts, automatically categorizing them based on learned patterns. This automation significantly reduces the need for manual data entry, allowing staff to focus on exception handling and more complex financial tasks.
The transition to AI-powered invoice processing has significantly improved operations across multiple teams. The new system can process a much larger percentage of invoices than before, drastically reducing the manual workload and freeing staff to focus on higher-value tasks. The imaging team, which handles initial triage, now sees fewer tickets as the majority of invoices are processed automatically. Additionally, the outgoing payments team receives invoices faster, minimizing delays and ensuring a smoother financial workflow.
This AI initiative, led by Doug Rose, Matt Dupie, and the Accounts Payable team, represents a major step forward in leveraging emerging technology to improve essential business functions at U-M. And they’re not stopping there—the Shared Services Center is also designing an AI-themed staff meeting to foster awareness, collaboration, and innovative thinking about AI’s role in their work.
SSC is enhancing efficiency by integrating AI into core processes and demonstrating how AI can be a strategic tool for continuous improvement at U-M.
To learn more about unique uses of U-M’s GenAI services visit the Use Cases page on the GenAI website.
