Nobody's Ignoring AI. The Calendar Is Full.

From where the AI program sits, the middle of the org looks like the slowest part. The strategy was communicated. The licenses went out. Inside the department, nothing starts. After the second or third nudge you get the same sentence back: "let's revisit this once things calm down."

We don't read that as a brush-off. It is an accurate description of where that person is standing. It is also a sentence that guarantees nothing will ever begin.

This is written both for people in that seat and for people trying to move someone in it. The assessment shows up near the end, but the structure comes first.

The load already concentrates there

Break down a manager's week and you find two sets of demands arriving from opposite directions. Leadership wants the transformation moving. The team says they are already at capacity on the day job. Layered on top of that: the numbers, the deadlines, the people stuff, whatever caught fire this morning, and a standing wall of meetings and cross-team coordination.

Add "and also drive AI adoption" to that, and the outcome is not in doubt. It goes to the bottom. Not out of apathy, but because more is arriving than can pass through. Read that as low commitment and the only remedy left is a motivational speech.

Our position: deferral here is a routing problem, not a character trait. When a junction is already over capacity, adding another item does not get it through faster. Something has to come off first.

No time to sharpen the axe

There is an old exchange where someone tells a person sawing at a tree that a sharper blade would cut faster, and gets back: "can't stop, too busy cutting." AI adoption is currently sitting exactly where that blade sits. Everyone agrees sharpening would help. Nobody can find the moment to stop.

One thing has to flip here. It is not "too busy to work on AI." It is working on AI in order to get less busy. Kept the first way round, the precondition for starting never arrives.

And the strongest early wins are usually not the big repetitive processes. They are the small scattered annoyances: low volume, high variation, never worth writing a procedure for and never worth automating. Reading something and summarizing it, or pulling up how a similar case was handled before, tends to be work that got skipped by every previous efficiency push. That is normally where the sharpening starts.

Room does not appear. It gets taken.

The calm month is not coming. Plan on its absence.

The largest decision available to a manager here is choosing what to stop doing and what to move first, without any spare room to do it in. Leave that decision open and AI adoption lands on the shelf marked important but not urgent, and slides quietly backwards from there.

Deferring is not neutral, either. "Next quarter" reaches the floor as a no. You may feel you left the question open; they heard a decision not to change. There is more on how that translation happens in the piece on what a meeting comment costs.

Split it four ways and you stop carrying all of it

"Just decide" is not actionable on its own. Splitting the manager's role into four pieces makes it workable.

• Decide what gets priority and what stops. Cut the number of open deferrals. This one cannot be handed to anyone else
• Clear time to try something, with the scope named. Narrow it to "two weeks, this one workflow"
• Connect leadership's language to the team's, push the team's blockers upward, and take on the cross-department coordination
• Protect the attempt until it produces something. Do not leave the person who tried standing on their own, and do not make a failed attempt their fault

Of those four, only Decide and Protect genuinely require the manager. Clear and Connect can go to someone in the team. If you are holding all four and saying you have no time, handing two of them over is the move that comes before anything else.

Don't pick who gets them by feel

So who gets them? Usually they go to whoever seems most into AI, or whoever looks least busy. Both are unreliable signals.

The person who should own Clear is not the person who knows the most about AI. It is the person who can actually move budget, time, and permission, so that the person experimenting is allowed to experiment.

The half of Connect that turns one good result into something repeatable belongs to a different person again.

The AI usage type assessment we build is 24 questions across 7 factors and 15 types, and it exists to answer exactly this question of who should get which piece. It takes about four minutes, costs nothing, and needs no signup. Have a few people on the team take it and line the results up: the roles that are covered and the ones nobody is holding become visible together. Once you can see the gap, the next request writes itself.

One honest limit. The assessment reads the shape of how someone works, not what drives them. For that, asking is faster. Not trying to catch motivation inside 24 multiple-choice answers is a deliberate design decision rather than a gap.

Move one thing today

No plan required. Before the day ends, settle exactly one of these.

• Pick one task in your own week and hand it to an AI tool once, as a test
• Take one tedious recurring task you have run for years and look at it again
• Name one meeting or one document that exists out of habit, and stop it
• Give the Clear role to someone on your team

What moves this is not the size of the plan, it is how quickly something gets started. Whether "sometime" becomes "one thing today" is the whole question. The deferral is not a character flaw, but leaving it deferred is a decision, and it is yours. So is the other one.