What Thousands of Questions Revealed About the Hidden of Value AI
Productivity
Mark Rosanes, AI Solutions Manager
AI Solutions
Tools
Shasta Chatbot [Ivy and Ocelot, by Gravyty]
Problem
University staff often spend significant time answering the same questions through phone calls, emails, and support requests. While each interaction is important, the volume of repetitive inquiries can reduce the time available for more complex customer needs.
Solution
Shasta helps by answering common questions using trusted university information, allowing staff to focus on higher value work. Over time, the conversation data also revealed valuable insights into customer needs and opportunities for service improvement.
When people think about AI, they often focus on what it can create. Our goal with Shasta, our AI virtual assistant, was simple: help users find answers to common questions. What we discovered along the way was that every conversation also provided insight into the services and information our community needs most.
At the University of Houston, Shasta has supported more than 72,000 conversations in its first 18 months. Most of those interactions were not complex. They were the same types of questions UH staff answer every day via phone or email: How do I enroll? What is my account balance? How do I reset my password?
Individually, these questions only take a few minutes to answer, however, collectively, they represent thousands of hours of staff time spent repeating information that already exists. This is a clear example of where AI can make a difference.
The obvious benefit to using a chatbot for customer support is that Shasta leverages trusted university sources to create answers to recurring, common questions without requiring staff intervention. That allows employees to focus on other work or step in when actual escalation to a person is required.
What I did not expect was the second benefit.
Every conversation creates trackable, measurable insights. Instead of relying on assumptions about what people need help with, we can see the actual questions being asked. We can identify service gaps to improve our information sources, and we can better understand where users encounter challenges.
One statistic stood out to me: about 40% of Shasta conversations occur outside normal business hours. That shows us another important insight: the need for support does not stop when offices close.
One of AI's greatest strengths is reducing repetitive tasks, so people can focus on the work that matters most, while giving leadership better visibility into what their communities actually need. Sometimes the biggest impact comes from Shasta answering the same question thousands of times so people don't have to.
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.