The current pressure around AI is easy to understand.
Managers are being asked where it can speed things up. Teams are being asked why they are not using it yet. Work that used to take days now looks, from the outside, like it should take minutes.
Some of that pressure is useful. Some of it needs a bit more thought.
The least interesting question about AI is whether it can make something faster.
It often can.
The better question is whether faster is what the situation needs.
A poor decision made quickly is still a poor decision. A weak analysis produced in seconds is still weak. A misunderstanding moved through an organisation at higher speed does not become less damaging because the tool was impressive.
AI is a force multiplier. That is the point. But it multiplies the quality of the thing it is attached to.
This is where the management conversation can go wrong.
If the instruction is simply use AI to speed this up, the tool gets pointed at whatever work is most visible: writing, summarising, tickets, analysis, documentation, code. Some of that will help. Some of it will just make bad inputs easier to produce.
The value is not in making every step faster.
The value is in using AI at the moments where better leverage changes the outcome: exploring options before a decision is made, testing assumptions before a plan hardens, summarising enough context that the right person can make a call, or turning rough thinking into something clear enough to challenge.
Used that way, AI does not replace judgement. It gives judgement more surface area.
AI does not remove the need for judgement.
It makes judgement matter more.
The teams that use AI well tend to be specific about where it belongs.
They do not ask how can we add AI to this process. They ask where in this process does a person need more context, more options, or a clearer first draft.
That framing changes the work.
AI becomes a tool in the hand of someone accountable for the result, not a magic layer sprinkled over a process nobody has understood properly. The human still decides what matters. The human still notices when the answer is plausible but wrong. The human still owns the trade-off.
The tool helps. Sometimes dramatically. But only when it is attached to someone who knows what good looks like.
So when a manager asks where AI can speed things up, I think the answer should start one step earlier.
Where are we making decisions with too little context?
Where are capable people spending time getting to the starting line?
Where would a better first pass change the quality of the conversation?
Where is someone using real brainpower on work that is predictable and repetitive?
Those are the places AI can matter.
Not because it makes the organisation think less, but because it can help the organisation think better, sooner.