“Faster" is the most expensive word in AI
By Polly Barnfield, OBE, CEO of Maybe*
Ask a business what they want from AI and, nine times out of ten, the first word out is "faster." Faster reports. Faster content. Faster responses. Faster everything. It is the instinctive answer; it sounds obviously good, and it is quietly leading a lot of companies in exactly the wrong direction.
I want to make a case that "faster" is the most expensive word in AI. Not because speed is bad, but because chasing it is how most businesses end up spending money on AI and changing nothing.
The seduction of faster
Faster is seductive because it is easy to imagine and easy to buy. You can picture it immediately: the thing that takes an hour now takes ten minutes. You can put it in a business case. You can demo it. Everyone nods, because who does not want their existing work to happen more quickly.
And you can get it, genuinely. Point AI at almost any task and it will help someone do that task faster. Drafting, summarising, searching, replying. All of it speeds up. So the promise delivers, in a narrow sense. People are faster. The demo was true.
Then a quarter goes by, and someone asks what all this AI actually changed, and the honest answer is: everyone is a bit quicker, and nothing is different. The reports still get written, just faster. The data still gets moved by hand, just faster. The same work happens, in the same shape, slightly compressed. You bought speed, and speed is exactly, and only, what you got.
That is the expensive part. Not the money, though there is that. The expensive part is that "faster" felt like progress, so it used up the appetite, the budget and the attention that could have gone toward the thing that actually moves a business. You did the comfortable version, it half-worked, and now everyone is a little less inclined to try the harder version, because did we not already "do AI"?
Faster keeps the work. The point is to remove it.
Here is the distinction underneath all of this.
Making a task faster keeps the task. Someone is still doing it. You have just shrunk how long it takes them. The job is still on their plate, still their responsibility, still a thing they have to remember and start and finish. You have optimised it, not removed it.
The businesses getting real value from AI are not optimising their tasks. They are removing them. They take a recurring job, the weekly report, the data entry, the follow-up, and they arrange for it to be done without a person, and handed back finished. The person does not do it faster. The person does not do it at all. It is gone from their week, and it stays gone.
That is a categorically different outcome, and it does not come from the word "faster." It comes from a different question entirely. Not "how do we speed this up," but "why is a person doing this at all."
The question that actually changes things
"Faster" asks: how do we do this quicker?
The more valuable question asks: does a human need to do this?
Those two questions send you in completely different directions. The first leads to a slightly optimised version of what you already have. The second leads to work leaving your business altogether, which is the only thing that shows up as real, structural change.
And the second question is uncomfortable, which is exactly why most companies avoid it and reach for "faster" instead. Asking whether a human needs to do something means admitting that some of what your good people spend their days on never needed them. It means changing how work flows, not just how fast it moves. It means some tasks disappearing, which raises awkward questions about roles and time and what people are actually for. "Faster" avoids all of that. It lets you keep everything exactly as it is, just quicker. That is its appeal, and its trap.
A simple way to catch yourself
Next time you or your team reach for AI to make something faster, pause on one question.
If this task could be done twice as fast, would that actually change anything, or would we just do the same thing with a bit of time left over?
If the honest answer is "we would just have a bit of time left over," you are optimising, and the gains will quietly evaporate into the general busyness of the week, the way saved time always does. That is the "faster" trap, closing.
But if you ask instead, "could this task simply not be ours any more," you are onto the thing that compounds. Because a task you remove does not give you a bit of time back this week. It gives you all of that time back, every week, permanently, and frees you to go and remove the next one.
Faster is a feature. Removal is a strategy.
None of this means speed is worthless. Faster is a perfectly good feature, and sometimes quicker really is all you need. The mistake is treating "faster" as the goal, because it caps out almost immediately. There is only so much value in doing your existing work more quickly, and you hit that ceiling fast.
Removing work has no such ceiling. Every job you take off a person is permanent, compounding, and clears the way for the next. One is a feature you buy. The other is a strategy you build.
So I would retire "faster" as the first word you reach for. Replace it with a harder, more valuable one: "gone." Not how do we make this quicker. What could simply not be ours to do any more.
That is the question that changes a business. "Faster" just makes the same business sweat a little less.
