Business Impacts
88%
Accuracy with identifying user intents
5,000+
Hours of manual effort saved
Customer Key Facts
- Location : University Park, Pennsylvania
- Industry : Higher Education
Problem Context
Academic advisors play a critical role in promoting students’ success. However, the rapid expansion of new programs, unconsolidated data, and time restrictions, makes it increasingly difficult for advisors to resolve students’ queries effectively. In addition, manual intervention is required for fetching data from multiple systems and interfaces, analyzing the data, and supporting the students.
Challenges
- Training the model on multiple intents across different topics
- Feedback looping via User Interface
- Integrating with on-premises databases
Technologies Used
Dialogflow
Google Cloud Platform
Python
Node.js
Scaling Student Advising Services With A Virtual Agent
Penn State World Campus, one of the largest universities in the United States with a student database that grew over 55% in a span of five years, wanted to scale its advising services. The full-time advisers collectively spent more than 5,000 hours a year assisting prospects, current students, and alumni across the globe with their day-to-day queries over email. During peak times of the semester, it took advisers longer to respond as they had to access different databases containing student information.
Solution
Quantiphi developed a virtual advising assistant using Dialogflow that resolves day-to-day email queries from 9,000+ students; sorting them into four common categories [i.e. change of campus, change of major, deferment, re-enrollment], automating the quick retrieval of relevant student data to run eligibility checks, and generating templated responses for live advisors to personalize; enabling live advisors to focus on addressing more complex requests. The virtual advising assistant was integrated with Penn State University's email system.
Results
- Quick request handling/faster response rate
- Reporting & metrics that empower data-driven decision making
- Greater cost savings
- Reduced manual effort - eliminated manual task of fetching data
