Reading the room before mentioning AI often separates the districts moving forward from those stalling out. Tech, budget, and timelines matter far less than investment in authentic leadership and culture, patterns Michigan Virtual's AI Strategists have observed again and again across the districts they support.

Based on conversations across seven Michigan districts and ISDs, and on patterns Michigan Virtual's AI Strategists have observed across many more, four leaders illustrate this in action:

• The Regional Coordinator: Watched superintendents in tight budget years unanimously vote to pay for Michigan Virtual's consulting support over free, piecemeal resources.

• The Curriculum Leader: Stepped into a leadership role when her superintendent stepped back, taking ownership of the district's AI task force, professional learning, and policy development.

• The Assistant Superintendent: A leader candidly owned her rollout missteps alongside her wins.

• The ISD Leader: Spent two years quietly building internal capacity before asking others to follow.

Start with a reason, not a tool

Before any of these districts talk about a tool, a policy, or a training day, they have each answered a more basic question: Why does this matter to us, specifically? Michigan Virtual's own approach is built around that question. Rather than arriving with a standard rollout plan, AI Strategists start by asking what a district already cares about, then look for ways AI can serve that interest rather than compete with it for attention. At Lapeer Community Schools, that meant tying AI training directly to an active district initiative aligned with FAME, a Michigan Department of Education (MDE) sponsored program that helps teachers collaborate and plan around the formative assessment process. Michigan Virtual's Delo worked with Lapeer's coaching team and other staff who were already leading that work, prompting them to think through how AI could support the formative-assessment training they were already doing, rather than teaching AI as a separate topic.

"Formative assessment is the focus, but there might be a tool [AI tool] to support that process."

— Karle Delo, AI Strategist Director, Michigan Virtual, on Lapeer Community Schools implementation

The same instinct shows up well beyond Lapeer. Muskegon Area Intermediate School District (MAISD)'s new five-year strategic plan, built with AI-assisted analysis of listening sessions across the county, doesn't mention AI once in its five goals: Partnerships, Wellbeing, Engaging Learning, Thriving Workforce, and Family & Community. AI accelerated the process of building the plan rather than becoming the point of the plan itself. St. Joseph Public Schools followed a similar logic when it introduced a new AI-focused course. District leadership framed it publicly in a November 2025 report as aligned with the district's BEAR strategic plan, and a timely response to the need for AI literacy, rather than a stand-alone initiative.

The through-line: districts that can name their own "why" for AI integration that is tied to a strategic plan, an assessment initiative, or an existing instructional priority, have something to anchor decisions to when the technology itself keeps changing. Districts chasing AI because it feels urgent, without that anchor, are more likely to drift toward the checkbox pattern described later in this piece.

What genuine engagement actually looks like

At Clare-Gladwin Regional Education Service District (RESD), the clearest leadership moment wasn't a speech or a mandate. It was a vote. Joe Trommater, Clare-Gladwin's Instructional Technology Integration Specialist and the region's REMC 5 (Regional Educational Media Center) director, walked into a first meeting with member-district superintendents and presented them with three choices, expecting them to choose the free route: piecing together resources from REMC and the Michigan Department of Education rather than paying for Michigan Virtual's outside consulting support. The member districts range widely in financial footing, some doing well, others under real strain, and getting them to agree on anything has historically been difficult.

But they voted the other way. Unanimously.

"I was actually pleasantly surprised that they went with the option that cost money. The superintendents in the room did the math and figured out what it would take to do it on our own…they said that once you divide it up among the districts, this was a very easy decision for them to make."

"That's when leadership first really showed up and said yes, we're willing to put skin in the game."

— Joe Trommater, REMC 5 Director / Instructional Technology Specialist, Clare-Gladwin RESD

That single decision, superintendents doing the math themselves rather than deferring the question, is what Trommater points to as the moment the region's AI work became real rather than aspirational. It also established a structural framework: a memorandum of understanding that spelled out each district's financial commitment and strategic AI consulting hours (both individually and for joint use), plus a governance model built on ranked-consensus voting rather than simple majority rule. Districts with very different risk tolerances and budgets stayed at the same table because no one could be steamrolled. That structure did as much to sustain participation as any single leader's enthusiasm.

At Lapeer Community Schools, Superintendent Matt Wandrie's leadership move was quieter than a vote, but structurally similar: he sat in on AI PD sessions alongside his own staff, and then intentionally handed the day-to-day work to someone well-equipped to run it. Michelle Bradford, Assistant Superintendent (Curriculum & Instruction), has driven the district's AI task force, professional development, and policy work ever since.

That handoff signaled something specific: what mattered wasn't Superintendent Wandrie becoming the district's AI expert. It was being visibly invested and placing clear ownership where it belonged. "He [Superintendent Matt Wandrie] uses AI, and he's very supportive. He attended the trainings when Karle (Michigan Virtual AI Strategist Director) was here, too," Bradford says. She describes him as thinking expansively about where AI could go and wanting AI coaches. In practice, he simply gives the team the go-ahead, and they run with it.

What followed was a task force that functions without requiring the superintendent to be in the room. Erin Koues, Lapeer's Software Systems Administrator, put it plainly when asked who's been driving the work: "Michelle's been driving it," she said, and then added, "in a way that everyone knows their job and what they're supposed to be working on."

That's the sustainability case: a structure that doesn't stall waiting on one person's bandwidth or technical expertise. Because ownership is clearly vested in someone empowered to use it, the work keeps moving even when the superintendent isn't necessarily the one pushing it forward.

At Muskegon Area ISD, Superintendent Randy Lindquist made a similar bet, played out over two years rather than a single moment: rather than leading every AI conversation himself, he deliberately handed the floor to others. A building principal led a discussion of new guidance with the admin team. A CTE director ran an activity showing how he personally uses Gemini. Department leaders across the ISD modeled AI tools in their own contexts before anyone else was asked to do the same.

"It's seeing different people and how they're using the tools, too, that I think has helped. It's not just 'sit and listen to a presentation,' but actually, 'here's a way it's being used practically in my job. Think about how you might be able to use that tool in your job.'"

— Randy Lindquist, Superintendent, Muskegon Area ISD

That sequencing signaled that comfort with AI needed to start with the people already in leadership roles, rather than being imposed as a mandate on everyone else at once. Lindquist prioritized getting his admin team to build their own knowledge first, on the theory that once leaders were genuinely using it themselves, that familiarity would spread naturally to their own staff and buildings.

What followed didn't require a large team. A handful of motivated, high-end users did most of the early work, not a district-wide push. That's the sustainability case here, too: a practice built on a few genuinely engaged people modeling real use, rather than on one leader's personal bandwidth or expertise, is the kind of thing that can keep running for years without burning out any single person, including Lindquist. It's also not something he describes building alone: when asked what he'd tell a superintendent wanting to move forward (or further) with AI integration work, his first recommendation is to lean on outside partners like Michigan Virtual rather than start from scratch.

Culture: not just the person at the top

Delo adds a dimension that's easy to miss if leadership is treated as a single yes-or-no question: culture. In her experience, districts with existing goals, clear structures, and a genuine tolerance for trying things that might not work tend to be better positioned, regardless of how invested any one leader is.

Trommater sees this play out across Clare-Gladwin's five member districts, which he describes as falling into roughly three camps: districts that are protective and cautious, districts that are enthusiastic early movers, and, the largest group at an estimated three-quarters, a cautious middle that's willing but wants to see it work first. None of the three, in his telling, is the "wrong" posture. He's also candid that the region has felt pressure to look further along than it is.

Alexandra Lee, an AI Strategist who joined Michigan Virtual this summer after working at an education technology company, complicates an assumption a reader might otherwise bring to this: that smaller districts move more slowly simply because they're smaller. Her hypothesis runs the other way. "It doesn't split based on district size," she says. Big districts can be genuinely innovative with a high tolerance for risk and change, she explains, or they can be bureaucratic and slow to move. Small districts operate within the same range; their barriers tend to be capacity, time, and resources rather than size itself.

She traces the difference to the leaders themselves as much as to the district. Some people are simply more excited by change than others, she says, and a lot of it comes down to change management and risk tolerance. But she's quick to flip the framing: trying something new carries risk, sure, but so does standing still, "because it's going to be used whether or not you're choosing to." Staff and students will use AI either way. The only real choice a district has is whether its leadership engages with that.

"There are uses for AI that will make you dumber, and there are uses that will make you smarter, and you have to know the difference."

"If you're going to push back against AI, make it for something real that you have all the data on, not because somebody told you it was going to drink all the world's water."

— Joe Trommater, Clare-Gladwin RESD

Trommater has been candid, too, about resisting the urge to move faster than his region can sustain. Having watched other early-adopter efforts get burned by moving too fast, he's explicit that Clare-Gladwin is "not going to be at the bleeding edge," even when that means possibly being seen, at statewide meetings, as dragging its feet. It's a leadership posture that trades the appearance of speed for something more durable. Once the region's task force work concluded, neighboring ISDs started calling Michigan Virtual to ask how to replicate what Clare-Gladwin had built. That outside validation is what Trommater points to when he says superintendents "went from feeling behind to feeling a little bit, not a lot, but a little bit, ahead of the curve."

Genuine, imperfect leadership vs. the checkbox problem

Michigan Virtual's AI Strategists are also candid about a pattern on the other end of the spectrum: districts that go through the motions of AI adoption (a policy document, a single PD session, a mention in a newsletter) without the underlying investment to back it up. Internally, the team calls it the "checkbox" problem^Footnote : activity that appears to be progress from the outside but doesn't reflect real engagement underneath. This pattern often emerges from good intentions combined with limited capacity, unclear ownership, or pressure to demonstrate quick progress. It's a pattern Michigan Virtual's AI Strategists look for most closely, since it's the hardest to see from the outside and the easiest for a district to fall into.

What's notable across these four stories is how differently the checkbox pattern shows up compared to genuine leadership. Genuine leadership doesn't mean getting everything right. It means being honest about what didn't work.

At St. Joseph Public Schools, Assistant Superintendent Amy Dirlam is candid about her district's first attempt at a district-wide AI rollout: guidance launched in August 2025, presented once at each building by the task force, and then the district largely stepped back for the semester. Reconvening the following January made clear that providing guidance wasn't enough. Staff needed reinforcement throughout the year.

"We all, as educators and leaders, have the ideas and the know-how of what needs to be done. It's the practical implementation piece."

"Nobody has all the answers yet."

— Amy Dirlam, Assistant Superintendent, St. Joseph Public Schools

Dirlam is equally direct about a gap she sees in her own leadership. Of all the groups in the district, she feels administrators have gotten the least of her direct attention, largely because limited bandwidth led her to lean on the task force as the primary carrier of the work rather than bringing building principals fully into it from the start. She addressed it by training department chairs, a strategy Michigan Virtual's Delo had specifically recommended: positioning them neither as classroom teachers nor as administrators, but as a middle layer who could model AI use in their own meetings first. It's not something she describes building in isolation, either: she credits the ongoing partnership with Michigan Virtual as part of what's made that progress possible.

"I don't ever want it to be like, 'Well, of course she would talk about that, because she's a techie,' although that is my background. But this has a curricular impact, whether we like it or not, it's a reality, and we've got to wrestle with it."

— Amy Dirlam, on why she chose not to lead the AI training herself

Genuine, imperfect leadership is the throughline across all four stories. None of it is about presenting a finished, polished narrative. Superintendents do the math out loud to figure out what makes the most sense; a district leader empowers his leadership team and trusts them to carry it out; an ISD superintendent quietly cedes the floor to his own staff; and an assistant superintendent names her own rollout missteps in the same breath as her wins. A checkbox is something you finish. Real leadership is something you keep doing.