Creating Service Insight Through Chatbot Conversations

If routine questions consume staff time or if you need better customer  insights,
keep reading.

Many UH departments face the same challenge: providing fast, consistent answers to common questions while ensuring staff remain available for the situations that require human expertise. Shasta, UH’s AI virtual assistant (uh.edu/shastabot), helps meet that need by providing support beyond traditional service hours while giving departments a better understanding of the questions and services their communities are seeking.

UH AI Solutions (AIS) offers Shasta as a service that enables departments to expand support availability while handling many of the routine questions commonly directed to staff using trusted university information sources. Since its launch, the virtual assistant has expanded to support nearly two dozen departments, reducing the volume of routine inquiries staff must handle while giving them greater visibility into the questions people ask most often.

THE CHALLENGE

University service teams receive repeated questions about enrollment, financial aid, account balances, technology support, and other routine needs. Every question matters to the person asking, but the volume and repetition can pull staff away from situations that require context or individualized support. While departments know users need help, they often do not systematically track the questions they receive, making it difficult to identify recurring needs and where additional support resources may be required.

THE SOLUTION

Shasta pairs AI with trusted university information to provide a highly available service option for common questions. Through integrations with campus technology systems, Shasta can provide personalized information such as account balances or enrollment details. Conversation data can also be leveraged to find content gaps and areas where information sources can be improved to offer better context.

“Shasta Chat Bot has become an important extension of our customer service strategy. It helps reduce repetitive inquiries by providing immediate answers to common questions while allowing our staff to focus on more complex support needs.” 
 
 - Kendrick Hagins, Director, IT Customer Services, Technology Services and Support

THE PROGRESS

Within the first 18 months, Shasta has been deployed across 23 service areas.

The data shows:

• Individual conversations become opportunities to understand user behavior and service demand patterns that may otherwise be difficult to measure
• Routine questions can be addressed using institutional information sources
• Customers expect access to support even when offices are closed

Even when measured against the lowest-cost support staffing model, a $15-per-hour student worker, the volume of conversations represents a significant amount of service capacity. At an estimated five minutes per interaction, 72,000 conversations equate to approximately 6,000 work hours, or $90,000 in labor-equivalent support activity over 18 months.

WHAT UH IS LEARNING

The most valuable lesson may be a better understanding of what people are trying to accomplish when they reach out for help and how the chatbot can transform support for participating departments.

WHY IT MATTERS FOR OTHER DEPARTMENTS

The broader opportunity is understanding how people seek information, what questions they ask most often, and learning where they encounter barriers. The long-term opportunity is not simply providing answers at scale but using those interactions to continuously improve the university experience.

THE TAKEAWAY

Shasta demonstrates that artificial intelligence can do more than automate support.  It can transform thousands of individual conversations into actionable data points to improve institutional knowledge. Departments can better understand customer needs and identify opportunities for improvement.