Six Tips for Leaders on AI Adoption

In just a few years, AI has entered the mainstream in offices everywhere, including on campuses. “It feels like everyone is talking about AI,” says Blair Thomson, senior international officer (SIO) and director of study abroad at the University of Mississippi. For SIOs, one challenge is separating the hype from real possibilities.

“You have to cut through the noise a bit,” says Thomson. “AI can and will affect recruitment, student advising, partnerships, and teaching. All of those things are leadership’s responsibilities. As SIOs, we need to understand the scope of AI. It requires both enthusiasm and caution.”

Because AI can encompass functions from routine communications to credential verification, the task of where to begin can seem daunting.

“Somebody described AI to me as a very small toddler,” says Wendy Lin-Cook, vice provost for enrollment management at Rensselaer Polytechnic Institute (RPI). “You wouldn’t ask the toddler to drive a car. But there are things that they can help with.”

Experienced leaders point to several guidelines that can provide a road map on how SIOs can successfully implement AI—as long as they understand what it can and can’t do well. Here are six tips to consider when implementing AI in the international office.

1. Identify the problem (or problems).

One benefit of adopting AI is that it requires offices to look at how they do their work—and how they can do things better or more efficiently.

“The process of considering AI is beneficial in and of itself,” says Kiki Caruson, senior international officer for the University of South Florida (USF) and vice president for USF World. Often, she says, office staff have been doing their work a certain way for a long time, and “it’s difficult to step back and say we could do it a different way and what would that mean.”

Taking a deep dive into office processes will identify where AI may be of use. “Start with the problem first and don’t be seduced by all the shiny AI tools out there,” says Thomson.

“Start with the problem first and don’t be seduced by all the shiny AI tools out there.” —Blair Thomson

Her staff came up with a series of issues to consider—and that are common to many international offices. “Younger people expect immediate answers, but we can’t be 24/7,” she says “Perspective students want to have accurate information about how to apply and who to email and get a general sense of what it is like to study here. Those were the kind of problems we have.”

RPI identified several common challenges, including increased applications and an avalanche of routine questions. Lin-Cook agrees that looking at the problems is the starting point, “It’s not, ‘I have the technology and how do I use it,’” she says. “Evaluation of the processes will help you see what the pain points are.”

2. Weigh the biggest issues with the potential benefits.

Because AI tools cover a wide range of workflows—everything from verifying applicants’ financial information to identifying prospective applicants—knowing where existing obstacles are will narrow down where AI might best work.

“When we think about the adoption of an AI tool, we’re thinking about process efficiency,” says Caruson. “Will it allow us to do some things we can’t because we’re so busy moving data from Cell A to Cell B?  It would be great to shift manual work to AI and use employees for higher-order work.”

One place many offices start is communication with students and potential applicants. “When we look at our communications, 80 percent are tier one questions with a very standard answer,” says Lin-Cook. “That is where we are leaning toward AI agents to help answer those questions and cede time back to the counselor who can answer more complicated or nuanced questions. Staff time is better spent having a more meaningful interaction with students instead of answering what are the 57 majors we have.”

“When we think about the adoption of an AI tool, we’re thinking about process efficiency.” —Kiki Caruson

Thomson agrees, based on her own experience. “I started as a study abroad adviser, and every 30 minutes a new student would come in with the same question,” she recalls. “We want AI to answer the repetitive questions that we get.”

Being able to get answers rapidly and at any time of day is simply better customer service for students. “Fifty percent of students who use the AI agent are reaching us during post-office hours,” says Lin-Cook. “It allows us to be much more present for the student when they need us.”

Relieving repetitive tasks will also help staff feel better about their jobs and aspects where the human touch remains essential. “AI allows my staff to do the work they want to do,” says Lin-Cook. “My staff took jobs in higher ed because they want to make an impact on the life of students. AI takes away the repetitive work so that we can have more meaningful engagement with students.”

Thomson advises offices considering AI to just focus on one or two opportunities. “Select a low-risk, high-value project,” she says. “What data can be used, what human help is required? Start small and systemically.”

3. Work with collaborators on campus.

Implementing AI isn’t just about applying technology to a problem. It also means grappling with issues of privacy, bias, and institutional risk.

“There’s a lot of groundwork you have to do to make implementation successful,” says Lin-Cook. “Partners like the CIO and the general counsel have to be involved at the beginning of the process. This tool is very new and powerful, so you have to be very careful there are no unforeseen consequences.”

For SIOs being approached by vendors, Thomson recommends including relevant offices on campus, like legal, the privacy office, and the access office, to understand the issues and the questions to ask.

When it comes to choosing a tool “Data, IT, legal—all of these people need to be in the conversation when you are with a vendor to make sure they are answering the right questions,” says Thomson.

“Partners like the CIO and the general counsel have to be involved at the beginning of the process. ... You have to be very careful there are no unforeseen consequences.” —Wendy Lin-Cook

USF formed an interdisciplinary AI Task Force, drawing upon people across the campus to tackle major issues. “It’s not just the IT folks,” says Caruson. “You need to have some kind of governance structure or task force around how the institution is going to navigate AI. This group can also function as a sounding board, because it’s a dynamic environment.”

Caruson says that having that structure in place is essential. “Without that, you need a compass,” she says. “I encourage participation in a work group, governance committee, or task force that will give you a sense of where the university is seeking to go and what its appetite for AI adoption and risk tolerance is.”

The conversation about AI should tie into the university’s overall vision. “Begin with the institutional purpose. We are trying to connect things we are doing back to the strategic plan, mission, and values,” says Thomson. “In global engagement, I ask: Does it improve access to the university, expand intercultural learning, and help us make better strategic decisions?”

4. Involve staff early.

As front-line workers, staff will experience the impact of AI first-hand. With stories about how AI can replace workers, they may be understandably leery.

“People come into implementation of AI technology with a preconceived notion of what it means for them,” says Lin-Cook. “You need to involve staff in the process and educate them.”

Being clear with employees from the beginning is essential for success. “Change management is part of the adoption of any new technology or operating system,” Caruson points out. “You have to manage people’s expectations by being very clear about why we’re making the choices we are. We are not seeking to replace people but to help people get out of super manual processes to do more critical thinking.”

“Change management is part of the adoption of any new technology or operating system.” —Kiki Caruson

Having employee feedback through the process is also crucial. “We allowed our employees to ask questions and allowed a free dialogue,” Caruson says. “It’s important to invite people into the discussion, especially those impacted by the adoption of a tool or process, to give them a voice in how effective that process is.”

Staff involvement assists in smooth implementation. “The staff loved being involved in the change process,” says Lin-Cook. “They see AI as a colleague that can potentially change the way they do their job and make it much more enjoyable.”

5. Remember that implementation and oversight take work.

Once an AI tool is chosen, adopting it into office workflows begins. While adding any new process or technology is work, operationalizing an AI tool is even more so. “We learned from the implementation process that training staff and the [AI] agent take more time than we allowed for,” says Lin-Cook.

The issue is that training an AI tool is an iterative process. “For the agent’s knowledge, you have to keep on providing feedback and a lot more guidance” in order to ensure accurate responses, says Lin-Cook. “It’s not something you could set and forget it.”

Her staff kept requesting the agent to repeat back to them how it made decisions. “That was very important for us to see to understand how AI was making judgments,” she says. “We have a very sophisticated staff with strong content knowledge to guide the AI agent, and that took a lot more time than we initially thought.”

“What I learned is that human capital investment in building AI is a lot higher than I thought.” Wendy Lin-Cook

Needless to say, there were glitches along the way. “It’s not surprising with any new implementation, there are hiccups,” says Lin-Cook. Initial communications with students weren’t clear and AI wasn’t always providing good information. “What I learned is that human capital investment in building AI is a lot higher than I thought,” says Lin-Cook. “If anything, we need more people than before to train the AI. I don’t think it’s a safe function independently without a lot of supervision.”

Blair says that she is constantly reviewing AI’s output for accuracy and risk. “I’m very meticulous about reviewing anything at large risk,” she says. “I apply controls appropriate for the risk. I look at [using AI] to brainstorm for an event as a different risk from answering an immigration question or communicating with a student in crisis.”

Ultimately, AI relieves people’s workloads, but it doesn’t eliminate them. “We do have to have a human at the end of this,” says Thomson. “You can’t say it’s human-regulated if all you do is hit a button.”

6. Keep an open mind.

Given how rapidly AI has established its usefulness in some areas, it’s here to stay. “It will be a very important tool for international staff,” says Lin-Cook. “As we train the AI tools, they will become 24/7 multi-lingual counselors.”

As a result, the profession itself will undergo change. “I do think international offices will move from being transmitters of information to strategic interpreters and advisors,” says Thomson.

Most of all, don’t be put off by the technology itself. “I would encourage folks to keep an open mind,” says Caruson. “You don’t have to be an expert in AI to appreciate what can be gained from it. Not every tool will be a quick fix, but you should remain open to what might help.”  •