It’s the end of first semester and a freshman college student is considering what classes to take in the spring. He knows he wants to declare his major – computer science – but isn’t yet sure what he wants to do for a career. Build websites? Create networks? Cybersecurity? Unaware of the pros and cons of each career path, he’s worried about whether he is choosing courses that will give him the right skills to get the job he wants. Until the higher education institute’s artificial intelligence bot texts his phone.
“Hello, my name is Renee. I’m here to help you as a success coach at the university. Add me to your phone and I’ll send you periodic updates, but you can contact me anytime you want to for help.” The student adds the contact as Renee, and a human-like relationship has been created with the bot. Not only can the student interact with Renee through text, but Renee can be contacted through Cortana, Skype, chat forms on registration pages, and more.
This isn’t theory, this is actually happening today. Renee is the name of the artificial intelligence bot created by Campus Management, and it is being tested in several higher education institutes across the country.
“We have this CRM platform and the number one problem facing schools is retention and persistence,” says Raymond Todd Blackwood, Vice President of Product Management for Campus Management. “So how can we have a communication platform that is fairly automatic, and it creates this relationship from a bot. With our CRM tool I’m able to have two-way SMS communication, I have canned responses, and I have a platform tool where I can build out communication plans that will outline all the key milestones that happen in a semester.”
Campus Management launched Renee years ago, but it wasn’t learning the way they needed it to. The sweet spot is in having it learn and understand intent, and not waste time configuring rules and canned responses. Campus Management chose Microsoft Azure as its cloud platform, and the APIs from Microsoft gave them the bot framework they needed – machine learning and streamed analytics. That gives the artificial intelligence the ability to learn to answer questions and learn to change the data it processes.
Processing Data to Make Course Recommendations
One of the largest challenges from an administrative standpoint is getting access to data. Typically ERP systems are locked away and hidden behind IT infrastructure. The only way to get data out is through static reports or asking the IT department for data. Today, data can be provided through APIs. CRM platforms give access to this information that can be queried through a web service. The artificial intelligence can then be used to process that data.
It’s the processing of data that gives the artificial intelligence program its power to change higher education. Renee is not only able to analyze data from the school, it can analyze data from outside sources as well.
Take, for example, market labor data about jobs. The artificial intelligence platform can pull in that outside data and analyze it to learn what skills employers are looking for in certain career areas. Then it takes information from inside the school to see which classes provide those skills. It makes recommendations not only on what courses will fulfill needs to graduate, but what courses will allow students to learn what they need to pursue the career they want.
“When you can mash up big data with local data at your institution and present it seamlessly to the student,” says Blackwood, “Now we’ve hit something that’s really special.”
This new guide will show you how to structure your RFP so you get the maximum return on investment for your budget.The Technology Manager's Guide to Creating a Technology RFP.
Breaking Off of Guided Pathways
Every school has guided pathways – sort of roadmaps that outline how a student would get through the school in a certain number of semesters while earning a specific degree, and allowing for some wiggle room with electives. The student can own their own journey, but stay within the guardrails of the pathways the higher education institute has created.
This configuration creates data that can be mined and analyzed for predictive analytics. The artificial intelligence agent then studies these pathways and learns from them so they can pick up where the school left off and create new, more unique pathways for students to follow. The artificial intelligence can create occupation insights that tell students what employers are looking for in certain career paths.
“If the school takes the time to match those skills with courses, students have a data point where they can see that if they take accounting they can be eligible for one type of career, but if they take a class in SAP they can be eligible for the same career but make $10,000 more in average salary,” says Blackwood.
By merging big data with student data and making it easy for administrators to configure it, the groundwork is laid to allow the artificial intelligence machine to learn. Then the AI can see how students apply to it, how the market changes, and continue to visualize the information for students or respond to students with answers.
Instead of administrators meeting with students and recommending courses, they help inform the artificial intelligence so that the AI can do so at a more granular and unique level to each student.
Artificial Intelligence and Higher Education Administration
If artificial intelligence is able to take all of this insight and offer suggestions for students on classes to take, can it use the same insight to offer suggestions to administrators on what classes to provide? Yes.
“We’re enhancing the pathways to put it in the hands of the students,” says Blackwood. “The registrar builds the path, the student enrolls in the program. They now get to see the path but make it their own – move classes around and choose electives. That data allows us to see what the demand from students is two or three years into the future. Our machine learning models can present to the registrar the classes they should offer in fall of 2020.”
It’s no different than any supply-and-demand model, and no different than any artificial intelligence that can provide business intelligence based on data. If more students are taking computer science courses year over year, then in several years you need more computer science classes to meet the demand. AI is able to analyze the data to provide these insights like never before.
This can be taken even further by predicting what grades students will get. The AI can predict grades based on past performance and other performance metrics for each class, and use that in the scoring model for at risk students and retention.
“You need to be careful in how much information you provide the students on the front end. Once they see the data they can take actions to manipulate it,” says Blackwood. “You want to be able to monitor, predict, test, and prove. You want that cycle to be ongoing so it learns and becomes more accurate over time.”
The artificial intelligence uses local and outside data to predict how many students will want to take a class. It uses the performance of similar students with similar academic histories to predict how each student will fare in the class. It takes the actual data of how each student performs in the class in order to refine its predictive model, and continues to get better through time. Now, the higher education institute has vetted information about demand, retention, and performance of every class to visualize for the administration.
Campus Management is still in the early stages of rolling Renee out nationally. There’s a lot left to learn, and only as the years go by will Renee learn from its mistakes and refine its predictive model to near-perfection. However, early adoption is promising and there’s no reason to believe Renee and other artificial intelligence won’t get even better over time. The trick is starting now so the AI can learn from your institution’s specific data sets.
Perhaps soon you’ll have as many guided pathways as there are students.