When people talk about artificial intelligence, they often focus on headline-grabbing possibilities: building software, generating complex graphics, or conducting research. Those uses are certainly impressive, but as an instructional designer at Tufts, I've found some of the most valuable applications of Microsoft Copilot are much less glamorous.

In fact, my favorite uses of Copilot involve tasks that most people might describe as tedious.

An example is comparing document versions. Faculty frequently share drafts of course materials, assignment instructions, and other teaching resources for review. Often, a revised version arrives with a note of, “I made a few small changes.” Sometimes the changes are shown in Word’s “Track Changes.” Sometimes they aren’t. 

When Track Changes hasn’t been used, finding edits can become an exercise in careful detective work. A paragraph may have been moved, or instructions might have been revised. A small wording change can significantly alter the meaning of an assignment or activity. Manually comparing documents is possible, but it requires time and careful attention, especially when differences are scattered throughout a long document.

This is where Copilot has become unexpectedly useful. Rather than reading two versions side-by-side and searching for every modification myself, I can ask Copilot to identify meaningful differences between the two. The results are often quite helpful, surfacing edits that might otherwise take me much longer to find.

Like many users, I've learned that success often depends on how I prompt the AI. Early on, I discovered that Copilot sometimes wanted to provide an exhaustive line-by-line comparison. While technically accurate, that level of detail wasn't what I needed. After some experimentation, I found that asking for substantive or significant changes generally produces more useful results. The process still requires human review, but Copilot dramatically reduces the time needed to locate potential differences.  

It’s also worth mentioning that I’m always mindful of data privacy when working with faculty content. That’s why I use Tufts’ instance of Copilot for working on these documents.

I've had a similarly helpful experience with Copilot when troubleshooting error messages. Before reaching out to colleagues in Tufts Technology Services, I like to investigate on my own whenever possible. Whether I'm trying to understand an unexpected error message, a feature that isn't behaving as expected, or a workflow that suddenly stopped, Copilot gives me a place to start.

What I appreciate most is not that AI always has the answer, but rather that it helps me ask better questions. Copilot can often suggest possible causes, identify settings worth checking, or point me toward relevant documentation. Sometimes the suggested solution works. Other times it doesn't. But even when a problem ultimately requires help from a human expert, I've usually gained a better understanding of the issue and can describe it more clearly.

These experiences have shaped the way I think about AI in instructional design. While it's easy to focus on transformative possibilities, there is also real value in helping people work through their myriad small tasks because collectively these take a significant amount of time and attention.

For teaching and learning professionals, that time matters. Every minute spent searching for changes in a document or hunting for a solution to a technical problem is a minute that can instead be spent designing learning experiences, supporting faculty, or working directly with students.

The future of AI at institutions like Tufts will certainly include ambitious and innovative applications. But in my daily work, some of the greatest benefits come from something much simpler: reducing the time spent on routine tasks. The work itself may be unremarkable. The time saved, however, is not. 

To learn more about using Copilot at Tufts, visit Microsoft Copilot: Generative AI Assistant