Turning Two Weeks of Work Into One Day of Review

Productivity

Our department often needs to review state requirements against university policies. The actual audit decisions are important, but much of the time was spent before those decisions could even be made. We need to compare new legislation against existing lengthy MAPP and SAM policies to ensure we are in compliance while also identifying where we are not then generating a report from what we’ve learned. It was repetitive, document-heavy work that could occupy nearly two weeks for a single review.

The focus was on reducing the time spent gathering and preparing information so the team could concentrate on evaluating results and assessing compliance. As I shared during a recent UH AI Solutions webinar, “We need to go out and search for tools that will help us in our work so that we can have more time allocated to other projects that will bring more value to the university.”

I began experimenting with AI using tools already available through Microsoft 365. The standard Copilot experience helped demonstrate what was possible, but for recurring compliance review work, I needed something more refined and repeatable. That led me to explore Copilot Studio and Power Automate, where I could build a workflow around the specific policies, statutes, and review processes our team works with on a regular basis.

Today, the process looks very different. The workflow retrieves and prepares documents, compares statutory requirements against university policies, and generates structured draft reports for review. Work that once required roughly two weeks can now be completed in about one day, and many of the initial comparisons can be generated in minutes. The biggest impact isn't simply moving faster. It's that time previously spent searching, organizing, and compiling information can now be redirected toward analysis and validation.

One of the biggest surprises was discovering how much could be accomplished without a technical background. I didn't start this project as a developer or automation specialist. Most of what I learned came through online research and testing ideas against a process I already understood well. Looking back, knowing the workflow was just as important as learning the technology itself.

The project also reinforced an important principle about AI adoption. We took a very familiar process and identified steps that are repetitive, time-consuming, and well-defined. In our case, much of the preparation work fit that description, making it a good candidate for automation.

At the same time, human review remains essential. Every AI-generated report is reviewed by an auditor before it moves forward. Drafts can be revised, rejected, or approved based on human evaluation. The workflow helps prepare information more quickly, but professional judgment and final decisions remain with the auditors.

What began as an experiment ultimately transformed a review process that once occupied nearly two weeks of effort into one that can be completed in about a day.  For anyone exploring AI, my advice is simple: start with a process you know well. If a significant portion of the work involves searching, comparing, organizing, or preparing information, there may be an opportunity to reclaim time and focus more attention on the work that matters most. 

_{The example, details, and outcomes described in this article were provided by the contributor. AI tools were used to help organize and draft the article, which was reviewed and validated by human staff to ensure accuracy, quality, and appropriate oversight.}