Almost none of them solidify the foundation necessary to do that work well. We call it Hold — the ability for individuals, teams, and organizations to carry the weight, stress, and complexity that change and uncertainty bring.
Trap #1 — Not Solidifying Your Foundation
The critical first step is solidifying individuals, teams, and the organization itself, so they can actually hold the transition, the turbulence, and the accelerated pace AI brings — before reaching for a single tool.
Trap #2 — The Momentum Trap
Someone in the room is pretty sure they don’t understand what agentic AI actually means for their team’s structure in two years — but they nod along anyway, because admitting uncertainty at this level feels like admitting weakness. Someone else has a real, reasonable fear about what this means for their role, and it never gets said out loud in the meeting — it gets said in the parking lot, or not at all. Everyone else appears to be moving, so standing still — even just long enough to think clearly — starts to feel like the riskiest option available. The instinct isn’t wrong. The trap is skipping straight past the one move that would actually make everything after it work.
Trap #3 — The Illusion of Performance
And then people start going rogue — usually with the best of intentions. They want to prove their worth, demonstrate real value, and not be the one left behind. Employees quietly adopting AI tools without sanction isn’t rebellion, and it isn’t laziness. It’s the individual version of a pattern this site already names elsewhere: carrying the load alone. Someone senses a shift nobody in leadership is addressing directly, feels unseen in it, and copes the only way available to them — alone, on their own initiative, without structure or support. When leadership notices the resulting productivity bump and rewards it without asking how it happened, something specific is happening: the organization is rewarding an illusion of performance — visible output, standing in for the shared, deliberate building real progress actually requires. An AI-performative culture is that exact failure, wearing new clothes: activity without structure, strategy, or focus.
None of this is a failure of intelligence or effort. Much of it is something more specific: people quietly protecting themselves in a moment when they’re genuinely unsure their role will still exist in its current form. Psychological safety alone doesn’t resolve that — you can build permission to speak honestly and still leave people afraid for their actual livelihood underneath it. A true holding culture requires both: the safety to say what’s true, and enough job security that saying it doesn’t feel like a risk to someone’s position. People need to know they are being held by the organization — not just permitted to speak within it.
Holding it well starts before any tool gets chosen. For a leader, it means being able to say “I don’t yet know what this changes for us” without it costing them credibility. For a team, it means enough safety — and enough security — that the real fear gets said in the room. For an organization, it means a small set of principles that stay fixed while everything else about AI shifts weekly: where human judgment is non-negotiable, what responsible use actually means here, what data governance requires.
Getting Hold right doesn’t finish the work — it makes the rest of it possible. Once people are genuinely held, Adapt has its own trap waiting: the temptation to borrow understanding from a vendor’s pitch deck instead of building it collectively, as a real, structural capability. And once an organization starts to move, Shape has a trap of its own too — mistaking a mandate for a rollout, and treating any given wave of AI as a project with an end date rather than a practice that keeps returning to Hold as the ground shifts again. More on both, next.
The organizations getting this right aren’t the ones with the best AI strategy document. They’re the ones where someone can still say “I’m not sure” out loud — and know their job isn’t what’s on the line for saying it.