A new national survey reveals that higher education leaders believe they have strong command of artificial intelligence (AI). But the administrators seem too confident. The survey shows that they don’t use the appropriate AI tools where institutions face the most pressure: in admissions and enrollment.

Liaison — an organization that helps more than 1,200 higher ed institutions use data-driven solutions to identify, recruit, and enroll best-fit students — surveyed presidents, provosts, deans, and directors across the country about their use of AI. Liaison today released its report.

The report reveals higher ed leaders’ blind spots on AI. It also points them to the underleveraged AI tools that will boost admissions and enrollment. These tools aren’t “nice to have.” They’re a necessity now that colleges close at a rate of one per week and an “enrollment cliff” looms.

Higher ed leaders’ confidence comes through clearly from the survey:

• 85% consider themselves extremely or very knowledgeable about how to use AI in admissions and enrollment management
• 81% say they’re already using AI with enrollment management
• 75% say they’re using AI in admissions
• 65% say leveraging AI for admissions and enrollment is “mission critical” or “very important”

But these confident self-assessments call to mind an AI researcher’s warning from more than 15 years ago: “By far the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.”

Dr. Mark Voortman, Liaison’s Chief Data Scientist, explains: “It seems clear from the survey that higher ed leaders understand AI tools like ChatGPT and chatbots. . . . But in reality, [these tools] are only a very small part of AI. The big problem is that institutions are missing out on other types of AI that have a much bigger impact on admissions and enrollment.”

Specifically, most institutions have failed to leverage predictive AI and prescriptive AI. Both apply machine-learning models. Predictive AI analyzes historical data to project what might happen in the future. Prescriptive AI recommends what actions to take to achieve a preferred outcome.

Instead, higher ed institutions rely primarily on conversational AI (chatbots).

• 73% of higher ed leaders use conversational AI for admissions . . . but only 41% employ predictive AI, and just 17% use prescriptive AI
• 75% use conversational AI for enrollment . . . but only 39% employ predictive AI, and just 20% use prescriptive AI
• Only 41% believe it will be “essential” to apply predictive AI to admissions and enrollment
• Just 31% say the same for prescriptive AI

Dr. Voortman notes that neither conversational nor generative AI will “really help you hit your enrollment numbers or get the class you want.” The report states the problem clearly: Institutions that don’t leverage predictive and prescriptive AI will miss out on students they should enroll.

The survey raises another red flag: 4 out of 5 higher ed leaders acknowledge that their data isn’t AI-ready. The trouble is, institutions can’t harness AI without a holistic view of their data. Information can’t be siloed in different systems. Machine-learning models need to see what actions have been effective in the past.

The Liaison report lays out practical ways to address all these challenges. It includes a Data Readiness Checklist any institution can follow. The report also shows the three steps higher ed leaders must take to optimize admissions and enrollment through AI — and to make sure their institution isn’t left behind.

For more insights on how higher education institutions use AI in admissions and enrollment, and how they need to evolve, download the full Liaison report.