CASE Currents’ November/December cover story, “Taking the Sci-Fi Out of AI,” was everything education advancement leaders expect—timely, insightful, well-researched, globally relevant, and skillfully written. As such, it did a great service in covering much of the landscape regarding how, where, and why institutions are increasingly using artificial intelligence in their strategies and departments.

The article also left a good deal of the AI landscape—and its potential use in advancement—unmapped. Perhaps that’s understandable; its scope and complexity are imposing.

While the content rightly focused on AI’s potential beneficial outcomes, it may have left some important distinctions and definitions unclear or unclarified. The very meaning of AI as conveyed in the piece was fairly limiting and therefore may have sewn confusion among institutional and department leaders, leading some to think they have robust AI tools and metrics when they may not. That may have let some believe they are doing all they can do—or setting the bar rather low when seen through a wider AI-definitional lens. 

The distinction between AI and “not AI,” by example, is not entirely academic. As is true in many other endeavors, knowing the right questions and asking them in the right way can be the difference between achievement and aspiration. Understanding what you are doing helps you understand what you may not be doing; seeing what you aim to accomplish can help you find the tools you need to get there.

The types of AI solutions richly described in “Taking the Sci-Fi Out of AI,” are narrow brush strokes in a more complicated fresco. Digital development officers and chatbots as described only tell part of the story of what AI can do.

To be clear, this is not dismissive of chatbots (as one example) or the many schools and departments that wisely use them. They can be, and are, force and reach multipliers that can open doors and engage targeted constituencies in ways humans may not be able to do routinely. The University of Illinois Foundation’s website, for example, has a chatbot specifically for donor engagement, available 24/7/365. That’s unquestionably beneficial.

By contrast, today’s AI platforms and implementations are nudgers—they do more than respond; they recommend. Likewise, they can open data that is too voluminous or too complex to be accessed and understood by mere mortals. And, again, they do not just put data in context but use it to advise and advance agency. Think of it this way: If your data analytics inform you that you’re in the top 15% of schools in, say, recent alumni engagement, that’s great. If it’s not telling you why that is or what specific actions you should take to crack the top 5%, it probably can and that’s where AI can help.

Amazon and Netflix are easy examples. Both use data to inform and benchmark—this TV series or tea set is ranked among the best in its category. But they also nudge—“try this.” That nudging is further along the four-part maturity curve of today’s AI technology:

• Definitive (telling what something is).
• Descriptive (helping you see it in context).
• Predictive (telling you what’s likely to happen next).
• Prescriptive (telling you what you should do now to make those things happen).

By and large, and with notable exceptions, advancement professionals (and the recent article) rest at step two. That leaves plenty of improving space before we bump our collective heads on the technology ceiling. More importantly, that gap leaves communities out, efficiencies unrealized, and advancements unmet.

As an aside—and perhaps to further clarify what AI is—linear regression and benchmarking, powerful as they are, do not do the same functions of other, more robust AI applications. Benchmarking, for example, is comparing data to other, often outside, data. And while it can often be informative, it’s not, at its core, predictive or path-provoking—definitional outcomes of AI.

The first step is recognizing that various forms of AI and technology exist that can help deepen engagements with constituent groups. The second step is asking which type of solution is best for the task at hand—driving philanthropic support (donor acquisition, retention, upgrading), increasing other forms of alumni engagement, creating more operational efficiencies—as well as organizational. I usually start by asking which muscles we’re not flexing and whether AI can help us do that. In many cases, it can.

One point made well in the Currents AI cover story was that institution and advancement experts need not be technology experts to ask, answer, and solve these questions. That’s true. Today’s technology is not sci-fi and help is abundant. All anyone really needs to be is an expert is their own situation and objectives—to know where they are and what they need to do.

Just as those needs and dynamics differ, so do the solutions. Not every objective can be met with new technology or the same technology. Other challenges are tailor-made for specific technology resolutions. Similarly, AI is not a magic letter combination, a panacea. It’s unlikely that “using AI” will solve anything by itself. Moreover, many people who think they are “using AI,” probably are not. Custom needs call for custom solutions. The good news is that the technology and AI landscapes are rich and varied and well-resourced, probably far more so than any single article can well convey.