If You Can Spot the AI, It's Not Working

"Good AI use disappears. Only the bad kind gets noticed."

There is a strange thing starting to happen in businesses getting good at AI.

You stop noticing it.

There’s no announcement that AI helped someone prepare for the meeting. Nobody particularly cares that someone talked a problem through with an AI assistant while out walking.

The proposal arrives. It is good. The meeting starts with better context. The follow-up happens faster. Someone catches something that would previously have been missed.

The work just gets done differently.

And increasingly, the time you really notice the AI is when somebody has used it badly.

Bad AI is obvious

We have all seen it.

The email twice too long that sounds like nobody you have ever worked with. The proposal full of polished words that somehow says almost nothing.

The presentation with thirty slides because producing thirty slides is now easier than simplifying it to ten. The meeting summary that went unchecked. The confident answer built on a fact that turns out to be fiction. You notice the AI because the work is worse. Somebody has used the technology to avoid doing the thinking, rather than helping them do the thinking better.

Good AI use is much harder to spot.

When it clicks

You can see the difference between someone who uses AI occasionally and someone who has built real capability.

At first, AI is a tool you have to remember to use. Could AI help me with this? Which tool should I use? How should I prompt it?

The shift happens when the person stops thinking about using AI, and starts reaching for it instinctively.

It could be something simple. Get these rough thoughts into shape. Challenge this email before I send it. Pull together the history before this meeting.

Or it could be a complex piece of work that takes an hour of back and forth, where the person knows how to structure the job, provide the right context, challenge what comes back and keep working until the output is good enough.

They're not using AI. They're just working.

Then it starts happening across the business

One capable person is useful. Multiply that by everyone in the building, and it's a different business.

People prepare differently. They get through difficult thinking faster. They carry more context into conversations. Routine work takes less effort. Good work gets finished faster. Problems get picked up earlier. And nobody necessarily attributes each of those improvements to AI.

That is where the apparent contradiction comes from.

The business may be using more AI than ever, but you start seeing less of the AI itself. It has disappeared into the work.

We have talked before about the difference between counting AI licences and looking at whether work has changed.

That’s still important, but this amalgamation feels like the next stage.

Invisible to whom?

There is an important distinction here.

AI becoming less visible in the work does not mean it should become invisible to the business.

A customer should not have to look at your proposal and think, "They've used AI to write this." They should think it is a good proposal. A colleague should not care that AI helped someone prepare for a meeting. They should notice that the person came prepared.

But the organisation still needs to know where AI is being used, what information it can access, what tools people are connecting and where human judgement sits.

Those are two different kinds of visibility.

As AI becomes more instinctive for the user, the organisation still needs the systems around it to maintain control. That becomes more important, not less, as AI disappears into normal work.

Look at your best users

If you want to know whether this is starting to happen in your business, look at the people who are furthest ahead.

What is different about how they work? Do they prepare differently? Can they move through difficult work faster? Do they communicate better? Do they carry more context? Do they finish work to a higher standard?

Take away the fact that AI was involved. Would you still notice the difference in their work? That's the real test.

Because the goal was never to create a business full of people who were visibly using AI.

It was to build enough capability that people knew instinctively when and how to use it, until eventually it became part of the job.

The technology should fade into the background. The improvement should not.

Find out if you'd notice where AI has already disappeared into the work

Most businesses can tell you when they rolled out an AI tool. Few can tell you how far it's actually spread, or where it's quietly connected to systems nobody's tracking.

The 5 Minute AI Readiness Snapshot helps you see how much visibility you really have into your own AI use - before it's too woven in to track.

Take the 5-Minute AI Readiness Snapshot

--

About the Author 

Mark Lucas is Co-founder and Managing Partner at Generation AI, where we help New Zealand business leaders build real AI capability.

Information provided is general in nature and does not constitute legal advice