Ten years ago, Georgia State University was facing a challenge that may seem strange on its surface. A surprisingly large number of admitted students — around 20 percent, all of whom had indicated their intention to enroll — weren’t showing up for the first day of class.
This phenomenon, which has come to be known as “summer melt,” wasn’t just an issue at Georgia State. Research suggests up to 40 percent of students from low-income communities who are accepted to college never matriculate. For many of those students, the maze of paperwork and other requirements — not to mention the fact that they are often the first in their families to attend college and may lack the sort of support so many college students take for granted — puts an end to their educational aspirations, before they even arrive on campus. It may go without saying that the pandemic has only exacerbated the challenge.
But the story of summer melt is not limited to higher education. In the world of work, there’s evidence that so-called “ghosting,” in which workers who have accepted a job stop engaging with their employer before their first day, has been on the rise. One study conducted in 2020 found 28 percent of job seekers report ghosting an employer, up from just 18 percent the previous year. And that was before the Great Resignation, which has seen more workers quit their jobs in a concentrated period than ever before.
While the names may be different, the challenge is similar: too many people aren’t taking the next step in their personal journey, be it educational or professional. What are the barriers standing in their way — and what can be done about them?
As it turns out, colleges and universities may have something to teach the business community. The same tools being used to address summer melt hold the potential to tackle ghosting in the workplace.
In 2016, Georgia State undertook a first-of-its-kind experiment: using artificial intelligence to help students navigate the transition to college. Accepted students were contacted via text message by a chatbot named Pounce (named for the school’s mascot), which proactively reminded them about key deadlines and provided guidance to make sure they completed all the requisite tasks. The results speak for themselves. According to a randomized controlled trial, students who received the outreach were 3.3 percentage points more likely to enroll in the fall. That represents hundreds, if not thousands, of students who owe their educational experience to the support of an AI chatbot.
Since that study, many more colleges have followed in Georgia State’s footsteps, as have states like Washington and Texas, and nonprofit organizations like the Common App. Using a framework called behavioral intelligence, which brings together 24/7 support with personal and contextual understanding, these partnerships are demonstrating the potential of AI to help students take the next step in their journey to and through college.
With these promising results, it stands to reason the approach may work for entry-level workers much the same way it does for incoming college students. What would that strategy look like in practice?
Over the past year, a growing number of employers have begun experimenting with the application of AI and behavioral intelligence to address ghosting and help smooth the path into a new job. So far, the results are promising. While the projects are just in their pilot phase now, they’ve demonstrated measurable decreases in ghosting and helped more early-career workers start their job on the right foot.
At a time when the so-called Great Resignation is upending the labor market as we know it, approaches such as these have never been more important as a way for employers to connect and communicate with incoming talent more effectively.
Just as importantly, AI is beginning to help employers listen at scale. Consider the results of one survey that a chatbot issued to jobseekers about why they dropped from the hiring process. For many respondents, logistics were the greatest challenge: one person missed an internet speed test and couldn’t reschedule and another received word that their background check never came back. These seemingly small things made the difference between starting a new job and starting over from scratch. And both are the exact sort of challenge that AI has helped to address in the higher education context.
While there’s still much more to learn, it’s clear AI can make a difference for people who need the right nudge, at the right time, to continue their progress, whether educational or professional. If these pilot programs in the workforce context prove effective, as they initially seem to be, how could chatbots play a role in reversing negative hiring trends and jump-starting the country’s economic recovery?