Using machine learning to transform online education and improve student outcomes

By using Google Cloud Platform's machine learning tools to understand and address student questions 24/7, Strayer University significantly improved support to its more than 48,000 students.

Strayer University’s student population is made up of mostly “non-traditional” students who face obstacles that the traditional student found on a campus quad does not. Most are working full time, they have children, and they often have previous attempts at higher education. Additionally, more than 80% of Strayer’s students take at least one of their classes online. The university recognizes that they cannot address online student issues in the same ways a traditional university might. It requires a unique approach.

"Deploying 'Irving' using Google Cloud Platform and Dialogflow has not only helped students get answers quicker, it’s also freed up University staff to focus on helping students with greater needs along their journey to complete their degree."

Joe Schaefer, Chief Technology and Innovation Officer, Strayer University

Introducing Irving: the friendly service agent

The institution already used a live chat platform with extended hours to accommodate working students. But it wanted to develop a virtual assistant that could learn by conversing with students and support them in real-time. Strayer leadership compiled a cross-departmental team made up of a behavioral psychologist, a software engineer, a full stack developer and a conversational designer. They sought out technology that would be up to the task and selected Google’s Dialogflow.

Daniel McCarthy, Executive Director of Artificial Intelligence & Automation, describes the language model of both Google Cloud and Dialogflow as “so robust that it’s mind-boggling how efficient and accurate it can be.” Google’s ease of use and flexibility, the ability to get up and running quickly, and the ability to transcend the chat bot model were important factors in deciding to partner with Google for this program. In less than a week the Strayer team created a prototype that could not only greet a student by name, but even know personal details like the student’s birthday. Named after the university’s founder, Dr. Irving Strayer, Irving could learn when a conversation was going off-topic and redirect the student to a live agent for further assistance. The change was remarkable.

Impact on students

Since spring 2017, there have been more than 163,697 interactions with Irving, serving over 20,000 unique students, with a 72% success rate in Irving's ability to understand student messages. Even students who don’t take online courses benefit as Irving is housed within iCampus, a tool all students have access to. Due to this success, Irving is now handling 74% of Strayer’s live chat volume–so student support staff can focus on those with more complicated needs. Reactions from students have been positive: when asked for feedback after interacting with Irving one student commented, “Very helpful, I can’t wait until I have another question!”

In the coming months, Strayer will leverage Dialogflow to build out sentiment analysis and further optimize the student experience. McCarthy emphasizes that Strayer is continually mining questions to add to its content and improve service. And, he adds, the university has higher ambitions for their technology: next the team plans to use the Dialogflow and Google Cloud technologies to identify students needing additional assistance and to offer tutoring and supplemental resources to help them better their life through education. McCarthy is optimistic: “Dialogflow's robust functionality fit right into the mission we were trying to accomplish. We are all very excited about this.”

Thanks for signing up!

Let us know more about your interests.