When We Talk About AI in Education, Here’s Just One Great Example

If your tech stack is still the same as it was twenty years ago, you’re missing out.

GUEST COLUMN | by Lance Hydrick

HARVARD.EDU

Whenever I land on a college website, I’m taken back to twenty years ago, when I was researching schools as a first-gen college student from a single-parent household. I was looking to transform my life, but I had no support, no network, no idea what I was doing, and websites that left me feeling helpless. Now, a generation later, it’s the same old experience. 

Today’s students—just like me 20 years ago—deserve better. 

They’re weighing higher education and private K-12 options against a flood of educational alternatives–and they rightfully expect an experience that works for them.

‘Today’s students .. deserve better. They’re weighing higher education and private K-12 options against a flood of educational alternatives–and they rightfully expect an experience that works for them.’

This is not a failing by enrollment departments. Enrollment professionals are thoughtful, innovative people. They’re just pulled in a million directions and haven’t had the technology to provide the personalized experience that potential students need–until now.

AI and the Evolution of Content Delivery

Everyone’s first instinct is to redesign their website, but it’s not like a new website has fewer populations to serve or less information to manage. Besides, who hasn’t already been there and done that.

Universities usually have around 8,000-9,000 pages on their sites; one school I work with has over 700,000 web pages. Compressing all that information into an easy-to-navigate website is just not practical.

Fortunately, with advances in AI, you don’t have to. AI can now quickly absorb all the information on a school’s website (even 700,000 pages worth), prioritize that information, and deliver it conversationally to curious prospects. The question is, now that we have this ability, how do we deploy it to build the best student experience?

How AI Makes Personalization Practical

AI has the processing power and flexibility to tailor interactions to each individual, which is a good thing considering that nearly 60% of students expect a personalized experience the moment they land on a school’s homepage. Where to place these interactions presents a dilemma: You could create a new interactive space to connect with students, but websites are crowded enough already.

Fortunately, school websites have standard built-in features that forward-thinking companies are efficiently optimizing for AI-powered personalization. And unlike traditional personalization that took weeks to implement and fine-tune, these tools can be configured and launched in minutes.

Site Search: From Conventional to Conversational

An ever-present website tool that’s long overdue for an AI-powered facelift is the infamously inefficient site search. 

Site search hasn’t really changed in the 20 years since I was researching as a student. But here’s the thing: it didn’t work then and it doesn’t work now. 

It’s a Google-lite, on-site wild goose chase that’s just as likely to deliver links from 2015 as it is to give people what they’re actually looking for. And yet, people use it.

‘It’s a Google-lite, on-site wild goose chase that’s just as likely to deliver links from 2015 as it is to give people what they’re actually looking for. And yet, people use it.’

3%-4% of website traffic still uses site search. For many schools, that represents thousands of visitors. Imagine how much that percentage would go up if site search actually worked.

Edtech companies are reimagining site search as an AI-driven conversation that delivers exactly what prospective students (and others related to the school), are looking for. With a knowledge base drawn from a school’s own web content, and the ability to communicate in natural language, we can now treat site search as a conversation that offers on-the-spot personalized information. Here’s what that experience can look like:

Natural language responses provide specific follow-up. Then links appear beneath–just like in conventional site search–to show where that information is coming from. With AI, the conversation can develop and pivot based on what the student needs. Also, since the school controls the content and web administrators can set creative parameters, the tool won’t hallucinate information like the average AI-powered large language platforms currently sweeping the tech landscape. 

Why This Works

With AI-powered tools, schools can personalize the student journey on their websites without rebuilding the whole experience for everyone else. They create an engagement layer that turns the impersonal, convoluted website experience into a tailored interaction.

Building this engagement layer into something like site search transforms a familiar tool into a personalized interaction engine. Schools don’t have to reinvent the wheel here. They just need to start finding ways to take tech-mediated elements of their interactions with prospects–RFIs, emails, search, etc.–and let AI turn them into a personalized conversation.

Researching schools shouldn’t feel the same as it did two decades ago, so it frustrates me to watch my kids start the same migraine-inducing process. They deserve better. They deserve schools that care enough to respect their time and curiosity. Thankfully, with this AI-evolution, we can finally give them that.

—

Lance Hydrick is founder and CEO of Halda.AI. Connect with him on LinkedIn.