There is a quiet assumption creeping into some businesses at the moment: If something can be automated, accelerated, or produced more cheaply with AI, then doing so must be progress...right?

I am not convinced that is true.

I have watched the industry long enough to see this pattern repeat itself. First with offshoring. Then with rigid process frameworks. Now with AI. Each time, the promise is efficiency. Each time, the hidden cost shows up somewhere else.

It's like playing Whac-A-Mole with your business.

AI absolutely has a place. A powerful one. Used well, it is a force multiplier. Used poorly, it is a shortcut that looks clever on a spreadsheet and quietly erodes value while doing little to positively influence the bottom line.

Where AI genuinely earns its keep

AI excels at work that is repetitive, time-consuming, or cognitively expensive without being judgement-heavy. Things like summarising information, analysing large data sets, generating first drafts, classifying inputs, or automating operational workflows.

In those contexts, speed matters more than originality. Accuracy matters more than taste. Consistency matters more than nuance.

That is why most teams report productivity gains almost immediately. Faster output. Lower effort. Less friction. There is nothing wrong with that. But even then, caution is needed to ensure the results are actually as accurate as the AI insists they are (that's a whole other topic).

The problem is when speed becomes the goal rather than the by-product.

That distinction matters because most teams report speed gains almost immediately. Industry research cited by WhatConverts found that over 90 percent of marketers say AI helps them generate content faster. Productivity goes up quickly. And herein lies the danger: assuming that faster output automatically means better outcomes.

The false economy of pure efficiency

Some of the most valuable work done in a business does not lend itself to efficiency.

Creative thinking is messy. Strategic thinking is non-linear. Bespoke problem solving rarely follows a neat path. These activities involve dead ends, disagreement, reframing, and judgement calls that only make sense in context.

When you ask AI to substitute for that work, you are not removing inefficiency. You are removing the conditions that allow quality to emerge.

Yes, you may get an answer faster. But you are far more likely to get something generic, derivative, and context-thin. Something that looks plausible but does not truly fit.

And this is where the false economy appears. Output goes up. Effectiveness quietly plateaus or declines.

AI has a tell. So does good creative and strategic thinking

After a certain point in your career, you develop an eye for quality. You can usually tell when something has come out of a strong creative studio or a genuinely sharp strategic mind. Not because it is flashy (although it can also be that), but because it feels considered. Specific. Grounded in reality.

AI-generated work has a tell as well. It is often smooth, confident, and oddly familiar. The edges are sanded off. The thinking is averaged out. It sounds right without being quite right. It's the creative or strategic equivalent of beige.

What is interesting is that even people who are not trained in creative or strategy still pick up on this difference. They might not be able to explain why something feels better or worse, but they respond to it all the same.

High-quality work carries signals of intent and care. Generic work carries signals of convenience.

Speed does not compensate for sameness

There is data to support this intuition. Research shows that while AI accelerates production, it also increases the risk of generic output. Nearly 40 percent of agency-side marketers surveyed reported that AI tools tend to produce repetitive or samey campaigns.

At the same time, the same research highlights a stark contrast: distinctive creative has been shown to drive many times more profit than generic advertising. In other words, the commercial upside comes from originality and relevance, not volume.

In other words, efficiency does not rescue ineffective thinking.

This is where many AI-first approaches stumble. They optimise for volume and velocity, then wonder why conversion rates flatten or engagement drops. The machine is doing exactly what it is designed to do. It is producing statistically likely outputs, not strategically differentiated ones.

A real-world reminder from customer service

We have seen this play out before, outside of creative work. Klarna is a useful recent example. The company moved aggressively to replace large parts of its customer service operation with AI, citing major efficiency gains and cost savings.

Not long after, they quietly reversed course and began rehiring human staff. Customer satisfaction had suffered. The savings were being offset by frustration, churn, and brand damage.

This is not a failure of AI. It is a failure of judgement.

The same pattern appeared years earlier with offshoring. Labour costs dropped. Resolution times increased. Misunderstandings multiplied. Eventually, many businesses realised the savings were being paid for elsewhere.

Efficiency always has a bill. You just don't see it immediately.

The opportunity cost no one measures

The hardest cost to quantify is the opportunity cost.

What ideas never surfaced because the brief was rushed through a prompt?
What strategic angles were missed because nobody sat with the problem long enough?
What differentiation was lost because the output sounded acceptable, so it shipped?

These losses do not appear neatly on a balance sheet. But over time, they show up in weaker brands, flatter growth, and harder-fought wins.

AI should amplify human thinking, not replace it

Used properly, AI can create space for better work. It can remove drudgery. It can surface patterns. It can accelerate exploration.

But the thinking still has to be done by people who understand the business, the customer, and the consequences of getting it wrong.

If your goal is to thrive rather than merely operate, the question is not whether you can use AI. It's where using it makes you better, and where it quietly makes you ineffective.

A final thought

Good digital work isn’t about doing more, faster. It’s about doing the right things well and in a manner that delivers the very best possible ROI for your business. That balance between efficiency and effectiveness is where some businesses struggle, especially as AI becomes easier to reach for.

If you’re wrestling with those trade-offs in your own digital platforms, websites, or internal systems, it’s often worth slowing down just enough to get the thinking right first.