Moving from predictive to prescriptive AI

More and more universities are adopting predictive analytics and forecast modeling to improve their recruiting and retention efforts. But what’s the best way to use those analytics and how can you tell if your implementation is off to a good start?eCN spoke with Jennifer Beyer, Vice President of CRM Product Management at Campus Management and former director of enrollment management at the University of South Florida Lakeland (USF), about how higher ed uses artificial intelligence (AI) in recruiting today and where it’s headed.eCN: It seems like AI is suddenly everywhere in higher ed. Is that your take as well?Beyer: When I left USF in 2013, AI use was not widespread, but over the last two year there have been a lot of these ‘how to use AI to solve problems’ stories. We are now seeing very real examples of how AI is starting to make an impact and become accessible to more institutions.eCN: Can you share some examples of what you’re seeing?Beyer: Using BOTs to support chat is one of the most common places where AI is leveraged. We are working with schools like Hope College in Michigan and Bemidji State in Minnesota to use conversational AI to automate responses through chat. Bots let colleges answer common questions and provide real-time answers to students at 8 am, 2 am, or 9 pm. In the next 18 to 24 months, chatbots in admissions will become more prevalent.We are also seeing AI used in analytics -helping to nudge students in the right direction. This is similar to the commercial experience they are already experiences. If Netflix can recommend shows to watch, shouldn’t we do the same for our students? This enables universities to go beyond using basic data points but instead to use AI that helps to understand sentiments and behaviors. This results in the ability to use data to feed into a model that is more prescriptive. For instance, we can use prescriptive data to look at signals—how a student is engaging with surveys and emails, how much time they are spending at the gym and in classes. Using AI, we can then recommend an action – for example text information about tutoring resources – to budge a student to get help in a class they are struggling in.eCN: What would AI help a college do with that kind of data?Beyer: In a nutshell, AI helps institutions to scale and deliver the right information at the right time. In a recruitment scenario, if you are someone who is goal driven and the college texted you information about internships and co-op opportunities for freshmen, that would make a much bigger impact than information about a sorority or fraternity. To be able to do that at scale, when managing caseloads of hundreds of recruits, is not possible today. At the same time, if I send you internship info and you respond with “Tell me more” or you put down an enrollment deposit, it gives me insight into your interests. AI also learns if you do nothing.Today, schools are spamming students because they have to cast a wide net. Leveraging AI helps us to personalize at scale, to drive our models so that they are responsive and always learning, and to help students find the school that fits best.