AI and automation can transform critical processes and provide insight into historical data.
AI and automation can transform critical processes and provide insight into historical data.
RSM created a document digitization solution for LSS of Minnesota, optimizing record management.
The innovative solution reduced discovery time and service request time from weeks to days.
Artificial intelligence solutions are making headlines with their ability to increase efficiency, enhance productivity and reduce costs. In some cases, these innovations can have a truly seismic effect on how an organization operates, transforming how critical processes are performed and providing more insight into historical data that was previously difficult to access, understand and manage.
Lutheran Social Service (LSS) of Minnesota is a 501(c)(3) nonprofit organization. Established in 1865, the organization supports a wide range of individuals, including youth and families, individuals with identified needs, and older adults. All told, 2,500 team members support 91,000 Minnesotans and serve with a vision that all people should have the opportunity to live and work in community with full and abundant lives.
Considering the organization’s long history, LSS of Minnesota has a significant number of historical documents. For one project, the organization needed to access information from around 100,000 files—millions of pages—spanning more than seven decades. They included handwritten letters, scanned forms and older formats originally stored on microfilm and later moved to Microsoft SharePoint.
Staff could not find the information they needed because files were not searchable or structured in any useful way; Microsoft Copilot estimated it would take one person 28.5 years just to read through the documents.
“We had thousands of scanned files in SharePoint,” says Phao Thor, application support specialist at LSS of Minnesota. “But they had no structure—no summaries, no search, no way to find what we needed. Some were blurry, some handwritten. People had to open many files just to find one answer. It took a long time, and sometimes we still came up empty.”
LSS of Minnesota needed a way to make that archive usable and turned to RSM US LLP for a new, more efficient solution.
Initially, RSM conducted a two-day workshop to demonstrate how Microsoft technologies could address LSS of Minnesota’s document management challenges and create a format that brought the organization’s data into a more consistent, accessible format.
Based on the findings from the workshop, RSM worked closely with LSS of Minnesota to develop a document digitization framework that combined the power of optical character recognition (OCR) and Microsoft automation and AI tools. The comprehensive solution converted the organization’s data into a raw digital format, enabled enhanced search capabilities, standardized the information, and created more streamlined, actionable results.
The RSM team leveraged Microsoft AI Builder for OCR document scanning, Microsoft Azure OpenAI for data summarization and Microsoft Power Automate to digitize and extract summaries from the millions of paper-based documents and records stored in Sharepoint. Many of the original files were almost illegible, with various forms of handwriting, and some documents had faded over time.
With support from the AI-enabled approach, LSS of Minnesota has experienced significant positive change in how critical historical information is accessed and presented.
“We were not chasing a trend—we were looking for a way to unlock what we already had,” says Heidi Leach, senior director of information technology at LSS of Minnesota. “Once we saw that AI could extract meaning from scanned documents—even handwritten or low-quality ones—it changed how we thought about our data. It was not just about managing an archive anymore. It was about reclaiming the stories, decisions and context that had been hidden for years.”
RSM also developed a conversational AI agent within Microsoft Copilot Studio to help internal and external users rapidly ask questions and get answers. Using Copilot Studio, the development of the AI agent took less than a week, with prompt engineering to determine the best output for the organization’s data. Through this process, the return was over 85% accurate, which was astounding to LSS of Minnesota leadership.
With a new framework in place, LSS of Minnesota has an effective mechanism to scan historical information internally and build a stronger data foundation.
“RSM and Microsoft gave us more than a prototype—they gave us the tools and guidance to continue the work on our own,” says Leach. “Since then, we have scanned around 100,000 files and built automated workflows that organize the data from every page. It became something sustainable that we could build on long term.”
LSS of Minnesota can now fulfill record requests much faster, reducing information discovery time and service request times from weeks to days. In addition, the organization’s employees can generate accurate reports for government compliance purposes with minimal effort, transforming regulatory processes from very challenging to automatic. LSS of Minnesota can now more effectively allocate their limited resources to mission-critical efforts by automating data searches and significantly reducing manual workloads.
“Our solution augments the work of the association’s employees,” said Ben Vollmer, a director with RSM US. “Humans still have to go through the files and validate the information, but we have cleaned up the data so they can find it quicker, use it better and focus more of their attention on pressing issues in the community.”
By implementing a more innovative and efficient records solution, LSS of Minnesota has more effective tools to accelerate their record management efforts. In addition, RSM’s document digitization framework is highly scalable and can also be adopted within other areas of the organization to further optimize key functions.
“We are seeing faster work, fewer mistakes and better decisions,” said Leach. “What was once an archive is now a source for information that helps us answer questions and support people better. We are still building on it—adding more structure, testing natural-language search and thinking about how to use it in other teams.”