Fluent. Confident. Wrong. In that order. Every cabinet meeting I've sat in since the second week of June has opened with some version of the same sentence: "The board wants to know our AI plan." Nobody's arguing AI matters. That fight's over. What almost nobody's asking is the question underneath it: the one that actually decides whether this fall helps your cabinet or quietly wrecks it. Before you put a machine on the payroll as your cabinet's newest ghostwriter, have you checked what it already believes about you? Here's what nobody tells a team of senior leaders walking into its first AI rollout: the tool doesn't create your team's judgment. It ghostwrites it, fluently, into every board memo and parent email and strategic plan you hand it, sounding exactly as confident in the ones that are wrong as the ones that are right. Two stories explain what's actually at stake this fall. One's old. One's brand new. Together, they're the whole problem. THE SENTENCE In third grade, a teacher tried to stop a boy from getting bullied. He'd been held back in reading, had migraines nobody diagnosed, and needed things repeated more than his classmates did. So she stepped in, pointed at him in front of the whole room, and said the thing she thought would protect him: "Leave that kid alone. That's the boy with the broken brain." She meant it kindly. He heard two words. Broken brain became the sentence he ran, silently, for the next decade: every time he wasn't picked for a team, every time he failed a test, every time he took the back-row seat hoping nobody would call on him. That kid grew up to be Jim Kwik, now one of the most recognized names in learning and memory training on the planet, the guy Fortune 500 companies and elite athletes call when they need to think faster under pressure. He told the story in a conversation earlier this year, and the line that stopped me wasn't the injury. It was what he said actually held him back: "Under stress, we don't rise to the level of our aspirations. We fall to the level of our stories, our systems, and our self-image." Not his intelligence. Not his skill. His self-image, inherited from a five-second sentence somebody else said, replaying under pressure for ten straight years. Now sit with the version your own cabinet is running. Maybe it's a board comment from three budget cycles ago that nobody's repeated out loud since, but everybody still quietly believes. Maybe it's a single paragraph an accreditation team wrote four years back that's still deciding what your provost thinks she's allowed to propose. Maybe it's simpler: a founding leader who once told this team they weren't ready for something, and nobody has tested whether that's still true. Here's the part that should actually worry you. Kwik didn't get better at learning by treating the injury. He got better by finding, and rewriting, the sentence. Your cabinet's talent isn't the variable in question here. It rarely is. The sentence running underneath the talent is. THE TURN Here's why this matters more this fall than any fall before it. For every year of your career until now, your cabinet's inherited sentence stayed mostly contained. It shaped a meeting here, a hesitation there, a proposal that quietly never got made. Slow damage, spread out over years. That containment is gone. Kwik, talking about the exact tools your board is asking about, put it plainly: "AI can scale and amplify. But it can only scale and amplify what a human brings to it." Feed it a clear, tested, collectively-owned judgment, and it multiplies your best thinking across every email, every plan, every decision your cabinet touches. Feed it an untested inherited sentence, and it does the exact same thing to that. (This is the exact gap THE TEAM INSTITUTE was built to close this fall: surfacing what the room already believes before the tool starts repeating it back at scale, which is different work than training your cabinet on the tool itself. Every vendor already offers that part. More on that in a moment.) TQ = IQ × EQ × PQ still holds. What's new is the multiplier. A cabinet with strong IQ and an untested self-image used to produce one flawed decision at a time. The same cabinet, with AI in the loop, can now produce that flawed decision a hundred different ways before lunch: in a strategic plan draft, a parent email, a board memo, and every version will read as confident, because the tool can't tell the difference between confidence and an old wound. THE OLD PLAYBOOK If you want proof this isn't just a Higher Performance Group observation, change management researchers are already redrawing the map. Kotter's 8-Step Change Model has run more transformations than any framework in modern management: create urgency, build a coalition, communicate the vision, empower people to act, anchor the gains. It's built for change that holds still long enough to sequence. A comparative model published this year in the Journal of AI, Robotics & Workplace Automation puts Kotter's eight steps next to a proposed AI Adoption Change Model, and the contrast is the whole argument in one picture. Source: Journal of AI, Robotics & Workplace Automation, Vol. 5 (2026), Henry Stewart Publications. Look at where each model starts. Kotter opens with urgency and a coalition, top-down, already decided. The AI model opens with exploring potential and discovering champions, bottom-up, still being tested. Kotter ends with anchoring the change in culture. The AI model ends with "cultivate AI-driven culture" and loops straight back into exploring again, because the tool never stops shipping new capability long enough for anything to get anchored. That's the structural version of what Kwik is describing at the individual level. AI adoption isn't a project you finish. It's a loop you run, and every pass through that loop amplifies whatever judgment your cabinet brought to it, tested or not. Which is exactly why the next five steps matter more than the eight above them. BUILD BEFORE YOU BORROW Kwik teaches a five-step discipline for exactly this problem, built for one person working with AI. It translates almost perfectly to a cabinet. You just have to run it as a room, not as one person with a laptop. Brainstorm first. Before anyone opens a tool, the cabinet talks through the problem out loud, together, with nothing feeding them an answer. This has nothing to do with whether the tool can be trusted. A room that's never practiced thinking together won't notice when the tool starts thinking for it instead. Understand the real problem. Most cabinets solve the presenting problem (enrollment's soft, a program's underperforming) instead of the one underneath it. Kwik's rule: spend the bulk of your time understanding what's actually broken before anyone reaches for a fix. AI is exceptionally good at answering the wrong question with total confidence. Imagine your own solution before you borrow one. Draft the cabinet's own answer first, even a rough one, before any tool sees the problem. This is the step almost every cabinet skips under deadline pressure, and it's the one that actually builds Team Intelligence. Skip it enough times, and the room stops being able to do it at all. Leverage the tool to stress-test, not to think for you. This is the one step where you deliberately fire the ghostwriter. Kwik's own practice: ask AI to argue against what the room already believes, to name the room's blind spots, to poke holes. Not "write our plan." "Tell us what's wrong with our plan." Different question, completely different outcome, and the only one of the five steps where you want the tool talking back instead of taking dictation. Decide, because the tool never has to live with it. AI has no stake in your decision. It doesn't sit in the parent meeting when the plan fails, or defend the budget line to the board. That accountability doesn't transfer, no matter how good the draft looks. Somebody in the room still has to own it. Five steps. None require new software. All five require your leaders to still be doing the thinking it's currently outsourcing, to the tool, or, just as often, to whichever one of you has quietly become the only person who does the real thinking before anyone else weighs in. (Kwik teaches this in far more depth than one newsletter can hold: his book Limitless and his Kwik Brain podcast go deep on the underlying discipline. No affiliation here, just work worth pointing you to.) THE CASE STUDY This is a composite, drawn from patterns across our work with 987 leadership teams , not one specific president's story. A university cabinet spent the spring drafting its first AI use policy the way most do: one dean built a draft using a chatbot over a weekend, and the cabinet spent forty minutes debating word choices in a document none of them had actually thought through together. The policy read confidently. Nobody in the room could say, if asked directly, what problem it was actually solving. Before the fall rollout, the cabinet ran Build Before You Borrow on that same policy, live, in one meeting. Brainstormed first, out loud, no laptops. Named the real problem: not "we need an AI policy" but "we don't yet agree on what we're willing to let a machine say on our behalf." Drafted their own answer as a room, in plain language, before any tool touched it. Only then brought a draft to AI, and only to argue against it. The resulting policy was four sentences shorter than the first one and, according to the cabinet, the first draft everyone in the room could actually explain from memory without looking at it. Same five people. One real conversation, finally had before the tool had it for them. THE FIRST MEETING You don't need a retreat for this. You need one meeting, early in the semester, run differently than the ones you've been running. Name the sentence (20 minutes). Ask your cabinet directly: what's something this institution believes about itself that nobody's tested in years? Write every answer down without editing them. You're not solving anything yet. You're finding out what you've all been quietly agreeing not to say. Run one live decision through Build Before You Borrow, out loud, before anyone opens a laptop (30 minutes). Pick a real one: the fall enrollment message, the AI policy itself, whatever's actually on this week's agenda. Brainstorm together first. Name the real problem. Draft your own answer as a room. Only then bring in a tool, and only to argue against what you just built. Agree, as a cabinet, on one rule for how AI gets used this year (10 minutes). Not a policy document. One sentence everyone says out loud and means: something like "nothing this tool wrote goes out first without a human draft underneath it." Whatever you land on, make it specific enough that you'd notice the day you broke it. If your first objection is that you don't have sixty extra minutes this month, you're already spending more than that reworking decisions that got made too fast, by whoever talked first in the room. This isn't additional time. It's the same time, spent before the mistake instead of after it. If your second objection is that your team will treat this as one more initiative, they're already operating under an initiative. It's just an unspoken one, and it's been running since before you took the job. THE MATURITY SHIFT Immature leaders think: "We need to move fast before we fall behind on AI." Mature leaders think: "We need to know what we actually believe before we teach a machine to say it for us." Immature leaders let the tool draft first and edit for tone. Mature leaders draft the room's own answer first and let the tool argue with it. Kwik calls the difference between these two postures victim mindset and heroic mindset: whether you let your hardest chapter keep narrating you, or you use it as fuel. It's a private decision about your own institution's story, one nobody else can make for you, and this fall, a machine is about to say that decision out loud, on schedule, to everyone. YOUR TURN Your cabinet is about to find out, in public, at scale, which story it's been running. No scandal required. A machine is about to say it back to everyone, fluently, on schedule, forever. Name it. What's the sentence your cabinet has been running that nobody's tested? Where did it come from, and is it even still true? You don't have to post the answer. But if you're willing to, I'd genuinely like to know what you find. If you already suspect which sentence it is, and you're wondering whether it's about to get amplified before you've even chosen a vendor: that's precisely what the TEAM INTELLIGENCE Assessment is built to surface. Fifteen minutes, no pitch, just a clear read on what your cabinet is actually working with before you put a machine in the room as your ghostwriter. higherperformancegroup.com/team-intelligence-assessment Across the 987 leadership teams we've studied, the cabinets that did this work before scaling anything (a new initiative, a new tool, a new plan) outperformed the ones that scaled first and diagnosed later, by a wide enough margin that "later" stopped being a viable strategy. The point is simple: make sure what gets sped up is actually worth repeating. If this is useful for someone else's first fall cabinet meeting, send it their way before their board asks the same question yours did. Tell them which sentence you found. Somebody in your network is about to let a machine ghostwrite a story they've never actually tested, and this might be the newsletter that stops them in time. Follow DR. JOE HILL and Higher Performance Group . — Dr. Joe Hill, Higher Performance Group | THE TEAM INSTITUTE