A new SMS-based chatbot is attempting to use natural language processing and machine learning to proactively engage at-risk students and connect them with on-campus resources. The company behind the solution, EdSights, is an education technology firm focused on improving student retention and wellness in higher ed.

EdSights believes that its providing universities with a sustainable way to "hear the voices of students at scale" and uncover insights on areas that are hard to measure, like a student's sense of belonging. The company reports that its chatbot has already engaged with more than 50,000 students at universities nationwide.

The latest from EdSights is the announcement that the company's student retention solution is now available to any higher education institutions throughout the U.S. An SMS-based, machine-learning powered chatbot, EdSights gathers real-time insights into college students’ experiences to identify those that may be at-risk for dropping out.

The platform promises to present colleges and universities with data-backed solutions that offer a better opportunity for intervention with at-risk students. EdSights uses natural language understanding to converse with students about their college experience and identify appropriate on-campus resources if a student is struggling.

"We programmed the chatbot to ask very direct, personal questions to pinpoint the exact reason a student is at-risk for dropping out and promptly connect them with the appropriate on-campus resource, while providing institutions with valuable data," says Carolina Recchi, co-CEO of EdSights.

EdSights then sends a text message to students at different points throughout a semester, for example to ask how they’re feeling about their upcoming midterm or final exam, or how they're balancing their course load while working full-time. Depending on the student’s response, they may be directed to an advisor or a specific set of resources that can better address the specific challenge.

If a student response triggers a certain crisis threshold, an alert will also be sent to university administrators to engage and intervene with the student directly. Students are also able to reach out to the chatbot proactively with any questions or concerns when in need of guidance.

EdSights monitors the trends in reported student struggles and outcomes to share with the university through an administrative platform. The admin platform displays data in a user friendly dashboard with specific "next steps" to help universities improve retention and student wellness.

“Administrators are given access to a dashboard that clusters students based on common struggles and provides insights into risks and retention rates," says Recchi. "We then use this data to inform administrators how they can intervene, whether there is a student who requires a follow up in the future or immediate intervention based on trigger words and phrases.”

The roster of companies that have turned their attention to student retention and success -- an issue that virtually every campus grapples with -- grows larger seemingly by the day.

EdSights' approach contends that relying on readily available data alone, like grades and classroom attendance, to identify at-risk students is not effective. The company no longer sees these as the key indicators, but rather symptoms of a larger problem. Moreover, the company believes this old way of identifying at-risk students leaves the students themselves entirely out of the conversation.

“The U.S. is facing a college dropout crisis and higher education institutions are looking for solutions in all the wrong places. Most colleges and universities operate under the assumption that what you see in the classroom, grades, and attendance, are the ultimate deciding factor of at-risk students, but these are often lagging indicators," says Recchi. "There are a lot of micro-barriers that students can be facing, which are often the root of the problem but are tough to measure. EdSights establishes an open dialogue outside of the classroom to identify this important, non-cognitive data early on and at scale by proactively engaging students and getting to the root of their challenge."