We are at the dawn of an artificial intelligence (AI) revolution and there is a lot of confusion about how this technology should be used to provide support for students, faculty and advisors in higher education. Terms like chatbots and natural language processing are thrown around, but many projects are not really AI and are not much better than the old IVR phone systems we hated. Press 1 for a better experience! Worse yet is that if you choose the wrong platform, you will be fighting a bad reputation and adoption issues for a long time to come. So let’s examine the top 10 features an effective digital assistant (or as some mistakenly term chatbot) should be offering your users.
1. Personalized answers, not “one answer fits all”
Any chatbot can answer an easy question where everyone gets the same answer such as, “What is a FAFSA form?” or “How do I reset my password?” In higher education, however, rarely are answers so generic. Higher ed culture is built on autonomy such that there are very few of these one-answer questions. Each department and school tends to have their way of doing things.
What if the user asks, “Do I need to sign up for a cap and gown for graduation?” The answer for a full-time student in the school of engineering will be different than the answer for a part-time student in liberal arts. If your chatbot can’t understand this difference, then it really is no better than pressing 1 on MovieFone.
2. True natural language processing, not just a crude search engine in disguise
A true AI-based digital assistant relies on AI modeling. Any chatbot that promises to crawl your existing web site and “just work” should be met with skepticism.
Time and again our student focus groups have told us they don’t trust or use search engines found on university owned web sites. The results are often contradictory and lack relevance. So why would your chatbot crawl a web site just as these web search crawlers do? You really are just deploying another web search disguised as a chatbot and that will not produce the experience or results you are looking for. A true enterprise digital assistant works in a fundamentally different and better way.
3. Deeply integrated with the SIS and LMS, and not just serving up static content
Any digital assistant needs to be able to help the student, and in order to help them effectively it needs to be able to talk to the two biggest systems of record in higher ed: the student information system (SIS) and the learning management system (LMS). Not only should the digital assistant be able to look up and use data from these systems, but also assist the user in conducting transactions such as paying their bill. If your chatbot is only providing links, than there is a missed opportunity much like we saw in the age of printed mapquest directions.
The digital assistant should have a catalog of adapters for your SIS and LMS systems, and a configurable suite of predefined intents like “Show me my grades” so you don’t have to build everything yourself.
Configurability is important, so you can change the delivered functionality based on your organizations needs, because in higher ed, nobody fits into generic boxes.
4. Understands the meaning of human language, not just a click through navigation tool
The promise of AI and natural language processing is that a user can type in anything and the bot will understand them just like a human would. Many people assume this is a given for all chatbots, but that is definitely not the case.
Crude chatbot solutions resort to providing a series of buttons to push you along a pre-determined path. Does that sound a little bit like the old IVR phone trees? Press 1 if you agree!
We believe a user should be talking/typing and not clicking. And the better the bot can understand humans (in over 100 languages too!), the less back and forth and the less clicking needed. Which, ultimately, leads to a faster, more enjoyable experience.
A true digital assistant is a navgiationless solution where people just “say” what they want to do.
5. Has the tools to allow you to grow its understanding, not a one-and-done solution
The whole point of artificial intelligence is that it gets better over time, right? How will your digital assistant get better unless it automates feedback and monitoring and allows you to model and test new training methods in multiple languages.
And the digital assistant provider should also deliver monitoring and training as a service. If your AI isn’t learning then it’s not AI.
6. Includes advisor and faculty “intents”, not focused only on students
There are a lot of student use cases that typically are catered for in early chatbot projects, but we find it a shame that faculty and advisors are often being ignored. After all, if you make the advisor experience more effective and comprehensive, the end result is that advisors can better service their students. Which, consequently, contributes to happier students who are graduating on time more often.
The challenge, however, is that your technology has to be up to the task. Securely automating advising processes isn’t a simple task. Data protection, advanced language understanding and entity recognition have to be the basis of this kind of solution. And many projects, using crude chatbots, don’t meet this standard.
Our digital assistants are being used by advisors to look up data on advisees based on PeopleSoft security settings, and even process tasks like advising notes, or approvals on course load. In one example, by using the digital assistant, a user can approve a course load in under 30 seconds, when previously it would require navigating and updating two web pages with about 10 fields and take multiple minutes (assuming they can remember to do it correctly).
7. Speaks the language of your institution, not just canned responses that could apply to anyone
While it sounds like a good idea to purchase a basic chatbot with pre-delivered answers to common questions, you have to consider what that really means. Not only will none of the answers be personalized to the user, but it also means answers will not reflect the nuances of how your institution’s processes work.
If one of your goals is to offer better service than human beings, 24×7 and in dozens of languages, than your digital assistant needs to provide better answers than a human would. Not canned responses that don’t fully answer the question.
When a vendor tells you, “You just plug it in and you are live in a week,” be skeptical.
The key is to have a business-user tool to allow you to easily, and without coding, manage and extend a catalog of questions. Additionally, you need a platform that isn’t generic by design. One that allows for sophisticated questions and answers.
8. Highly secure and protects sensitive data, not an open invitation for data breaches
In the age of cloud applications and data breaches, security is more important than ever. We have seen institutions deploy a non-authenticated bot for a narrow use case and believe that data security is not a risk. Even if your chatbot can only answer basic questions about admissions, you can’t prevent the user from typing in a social security number, a test score or other sensitive data. If you are not using an enterprise digital assistant, where is that data stored? Is it secure?
Further, the security of the bot channel is often overlooked. Channels like Facebook and SMS are used, but keep in mind, those channels are not secure, and there are no guarantees for how that data is stored or how it is protected
Using a secure channel like a web channel or MS Teams will protect your data and provide your institution the controls needed.
IntraSee uses the most secure cloud in the Enterprise: the Oracle Cloud. And for customers using Azure or Oracle Cloud for other enterprise data, there are even greater assurances.
We have also built in features like abbreviated logging of conversations to protect sensitive data ensuring that we are never storing your sensitive data in the cloud. This type of data security is something that basic chatbot vendors almost never can deliver.
9. Can handle an extremely broad range of questions, not a one-trick pony
Many chatbots live on their own islands. They are locked into narrow functionality specific to the service desk software they come with, or a particular vertical space like admissions or financial aid. As Amazon has taught us, people want a one-stop shop to serve all their needs.
Imagine a student has a question, but first needs to figure out which of the 20 chatbots they need to go to. That is a recipe for failure and isn’t really achieving the promise of artificial intelligence.
High performing AI should be able to walk and chew gum at the same time, just like a human! It is important to handle a wide variety of student questions, but it is also important you can leverage delivered skills from enterprise software partners. IntraSee’s digital assistant can combine our catalog of skills, delivered Oracle or PeopleSoft skills and custom client skills into that perfect one-stop digital assistant.
10. Helps drive student success, not a crude FAQ machine
It is one thing to be able to answer questions 24×7, but the goal ultimately is to help students. The best form of help is proactive, timely help.
A digital assistant should be capable of reminding students about their to dos and holds, as well as initiating a chat with key reminders that an assignment is coming up or their enrollment window is opening soon.
While reducing the cost of human-based service desks is the easy ROI case for digital assistants, their role in student success can be much more impactful.
If you haven’t seen a digital assistant or chatbot than can do all of these things, then just reach out to us below and we will setup a demo to show you live!