Responsible AI Implementation Lessons: What We Can Learn From The Leading Community Colleges

Across higher education, conversations about artificial intelligence are rapidly shifting from possibility to implementation. Yet the institutions making the greatest progress are not starting with the technology itself. They are focusing on a more important set of questions: Where can AI meaningfully improve the student experience? How should student data be governed and protected? Who should be involved in evaluating and deploying new technologies? And how can colleges innovate responsibly without allowing urgency to outpace sound judgment?

Those questions were at the center of two recent panel discussions with Brian Sailer, Executive Director of Student Persistence and Completion at Central New Mexico Community College, and Emery Peck, Vice Chancellor and Chief Operating Officer at the Ivy Tech Community College Muncie campus.

While their institutions serve different student populations, both leaders described a similar approach to AI adoption. At Central New Mexico (CNM), the focus has been on helping prospective and dual credit students make more informed decisions about allied health pathways. At Ivy Tech Muncie, the emphasis has been on helping dual credit students better understand how their high school experiences connect to future college programs and career pathways.

The most important lesson was not that both institutions identified the same use case. It was that neither began with the technology. They began with a clearly defined student challenge, aligned stakeholders around a shared outcome, and then evaluated where AI could responsibly add value. In many ways, that distinction may be one of the most important predictors of successful AI implementation in higher education.

Q1: Where should an institution begin?

Begin with the resource friction or student challenge, not the feature list.

At Central New Mexico Community College (CNM), faculty in allied health observed a recurring challenge: students were entering highly structured programs, only to later realize the field was not the right fit for them. By that point, changing direction often meant significant delays, added cost, and lost time for the student. This led to a reframing of the core question: Could students explore allied health pathways earlier, and make more informed decisions before committing to a program of study?

At Ivy Tech Muncie, the challenge was continuity. Students were successfully completing dual credit and dual enrollment coursework in high school, but many did not fully understand how those credits connected to a broader academic program, a career pathway, or their decision to enroll full time after graduation.

Neither institution began with the question, “How can we use AI?” They began with a more fundamental question: “Where are students getting stuck?” That distinction may seem subtle, but it fundamentally shapes implementation. It clarifies the audience, defines the problem to be solved, and creates a clear basis for evaluating whether the solution is actually improving the student experience. It also sets the foundation for designing the first responsible AI experiment.

Q2: Is it better to start broadly so more students can benefit?

Starting small is sometimes interpreted as a lack of vision. In practice, it is often a sign of disciplined leadership. A focused pilot gives institutions space to test assumptions, surface gaps in data, understand the workflows around the technology, and build internal trust before expanding to scale.

At CNM, the initial implementation focused on allied health pathways. Faculty and academic leaders were invited to engage with the experience as students would. Early testing surfaced important gaps, including inaccurate recommendations tied to programs CNM did not offer. These iterations helped the team develop a clearer, more grounded understanding of what responsible implementation requires in practice.

At Ivy Tech Muncie, the team reached a similar conclusion from a different starting point. An early concept aimed for a broader rollout across multiple student groups, but scope quickly became a challenge. As more audiences and outcomes were added, it became harder to define what the initiative was actually trying to achieve. Progress accelerated only after the focus narrowed to a specific dual enrollment population and a committed partner school.

As Emery Peck summarized: solve for one context first, then scale with confidence to many.

Q3: What makes AI implementation responsible?

Across both institutions, a consistent set of leadership questions emerged. Senior leaders did not need deep technical fluency to engage meaningfully with AI, but they did need clarity on ownership, data flows, accountability, and system boundaries. Who owns the information? Where do the answers come from? What data is truly necessary? Who can correct errors? And what happens when the system reaches the limits of its intended use?

At CNM, early users included prospective and dual credit students, many still in high school. That immediately surfaced questions around data access, accountability, and oversight. The team needed clarity on what information students would be asked to provide, how that information would be used, and whether staff could review both student interactions and the responses generated by the system.

As Brian Sailer shared "We’re using this as a way of increasing the size of our funnel into our enrollment pathways. Students get an idea of what they want to do and then we share that information with our outreach team .... And when I ran into mistakes with the platform recommendations, I was able to contact the team, and they were very quick to fix it and made sure the information was correct. I can’t do that with any of the other ones."

At Ivy Tech Muncie, Vice Chancellor and COO Emery Peck was navigating a related but distinct challenge. The college was eager to move forward while the broader institution continued refining its approach to AI and data governance. This led to a more constrained approach, exploring targeted, lower-risk applications that could deliver value without requiring large-scale movement of student data into external systems.

Q4: Does AI reduce the role of faculty, advisors, and counselors?

This is where many technology rollouts weaken. Access is mistaken for adoption. A platform is launched, a link is shared, and student use becomes the measure of whether the idea worked. Both institutions are taking a more grounded approach. They are placing the technology inside an existing network of relationships and making sure the people within that network understand its purpose.

At CNM, career exploration helps a student arrive with an initial sense of direction. An advisor or faculty member still helps the student interpret that information, understand program realities, and decide what comes next. The technology creates a starting point. The human conversation turns that starting point into context and next steps.

Ivy Tech Muncie’s adoption plan is equally dependent on people. Its initial users are in high schools, which means their first question about the platform is more likely to go to a counselor, teacher, or dual credit instructor than to an Ivy Tech administrator. The college is therefore preparing the professionals students already trust to understand the resource and respond when students need help.

Q5: What should colleges look for in an external AI partner?

The clearest answer from both conversations was not a specific technical capability. It was the ability to listen, collaborate, and adapt.

As colleges evaluate AI partners, features and functionality matter, but successful implementations often depend on something more fundamental: whether a partner is willing to understand the institution's goals, work within its constraints, and be transparent about what the technology can and cannot do. In an environment where many AI solutions promise transformational outcomes, that distinction matters.

At Ivy Tech Muncie, Emery Peck reflected on the college's experience working with several AI companies. What stood out was not who had the most ambitious vision, but who was willing to engage in an iterative process. As he noted, some providers approached the conversation believing they already had the answer. They were less interested in understanding the institution's specific challenges and less willing to adapt their approach based on feedback.

By contrast, the most productive partnerships are built through collaboration. Rather than beginning with a predetermined solution, the work focused on defining the problem, testing assumptions, gathering feedback, and refining the experience over time.

That pattern was evident across both CNM and Ivy Tech Muncie. The implementation process involved validating pathways, correcting institutional information, narrowing use cases, refining workflows, and continuously engaging the faculty, staff, and leaders closest to students. The technology was important, but the ongoing partnership played an equally important role in determining whether the implementation delivered meaningful value.

For institutional leaders, the lesson is straightforward: evaluate not only the product, but also the partner. The strongest AI implementations are rarely built through technology alone. They emerge from partnerships grounded in transparency, shared problem-solving, and a willingness to learn alongside the institution.

Q6: What does implementation require after the platform is selected?

Selecting a platform is an important milestone, but it is only the beginning. The real work begins when institutions determine how the technology will fit within existing student experiences, workflows, and support structures.

Across both conversations, leaders consistently returned to a set of operational questions that are easy to overlook during procurement. When will students encounter the platform? Who will introduce it? What should happen after a student completes an exploration activity? How will advisors, recruiters, counselors, faculty, and staff build upon what students learn? Who is responsible for monitoring accuracy, supporting external partners, and evaluating outcomes?

These questions are not secondary considerations. In many ways, they are the implementation itself.

At Central New Mexico Community College (CNM), career exploration is being connected to broader enrollment, advising, and student support efforts. At Ivy Tech Muncie, the work is closely tied to dual credit partnerships, high school engagement, and pathways into postsecondary education. In both cases, the goal is not to create a new student journey, but to strengthen an existing one.

That distinction is important. Successful AI implementations are rarely standalone technology projects. They are institutional initiatives that require alignment across people, processes, and technology. When those elements move together, AI can enhance the work already being done. When they do not, even the most promising technology struggles to create meaningful impact.

Q7: How should leaders define value?

For many institutions, conversations about new technology quickly turn to return on investment. While measurable outcomes remain essential, Emery Peck introduced a useful distinction: the difference between return on investment and value on investment.

Colleges should absolutely establish clear measures of success. Depending on the use case, those metrics may include enrollment, matriculation, persistence, pathway completion, student engagement, or staff efficiency.

At the same time, some of the most important indicators emerge before long-term outcomes are fully visible. Early implementation can reveal whether students are asking better questions, whether advisors and staff have richer context for conversations, whether external partners remain engaged, whether institutional information becomes more accurate, and whether the initiative is addressing a problem the community genuinely needs solved.

These signals should not lower the accountability standards. Rather, they provide a broader view of progress during the early stages of implementation and allows for timely adaptation of efforts. Focusing exclusively on the quickest available metric can cause institutions to overlook meaningful evidence of learning, adoption, and value creation.

For leaders, the challenge is not choosing between outcomes and impact. It is understanding how both contribute to long-term success. The strongest implementations establish clear performance measures while also recognizing the incremental gains that build trust, improve decision-making, and create the conditions for lasting institutional value.

Q8: What is the most important lesson for college leaders?

Responsible AI implementation should not be a race to appear innovative.

Across both conversations, one theme emerged consistently: successful AI adoption is less about moving quickly and more about moving intentionally. While CNM and Ivy Tech are advancing their AI initiatives, neither institution is treating innovation as an end in itself.

Instead, their leaders are following a disciplined approach. They are defining a meaningful student challenge, narrowing the initial use case, engaging the right stakeholders, testing assumptions, validating information, and establishing the governance and support structures needed for long-term success. Most importantly, they are using technology to strengthen, not replace, the human relationships at the center of education.

That is what responsible implementation looks like in practice.

It is not defined by the size of a launch, or the ambition of a promise. It is reflected in the quality of the questions leaders are willing to ask before asking students, faculty, staff, and communities to place their trust in a new technology.