AI changed the cost of making. Not the cost of being wrong

A few years ago, a campaign idea needed a strategist, a copywriter, a designer, a production budget and two weeks. Today a junior marketer can produce forty versions of it before lunch, in two languages, resized for every platform.

That is a real revolution, and I am not going to pretend otherwise. But I want to ask a different question from the one most people are asking. Not what can AI make? It can make almost anything. The question is: now that making is cheap, what is expensive?

Production got cheap. Being wrong did not.

The numbers on adoption are clear. The CMO Survey found that generative AI was used in 15.1% of marketing activities in spring 2025, more than double the 7.0% of a year earlier. In Gartner's 2025 CMO Spend Survey, marketing leaders said their generative AI investments were paying back through better time efficiency (49%) and cost efficiency (40%).

So production is faster and cheaper. But look at what has not changed. A wrong positioning still costs a year. A message that misreads the culture still costs a reputation. A brand that sounds like everyone else still costs you every campaign that follows. AI lowered the price of making things. It did nothing to the price of making the wrong thing.

AI changed the cost of making. It did not change the cost of being wrong.

The sameness problem

Here is the risk few teams are discussing. When every brand uses the same handful of models, trained on the same internet, prompted in similar ways, the output starts to converge. The same structures, the same phrases, the same polished, forgettable images.

In a feed where most ads get less than two and a half seconds of attention, sounding like everyone else is the most expensive mistake you can make. The brands that win will use AI to go faster, and their own judgement to stay different.

Trust has not caught up

There is also the audience to think about. The global study on trust in AI by KPMG and the University of Melbourne, covering more than 48,000 people in 47 countries, found that 66% of people use AI regularly, but only 46% are willing to trust AI systems. Among employees, two-thirds (66%) say they have used AI output in their work without evaluating it, and 56% say they have made mistakes in their work because of AI.

Put those numbers together and you have a warning for every marketing team. Your audience is wary of AI. Your own people are over-relying on it. That combination is where brand mistakes are made.

Three lists every team should write

The answer is not to slow down. It is to be explicit. Every marketing team should write three short lists and put them where everyone can see them.

• What AI does. Research summaries, first drafts, versioning, resizing, translation drafts, testing variations, reporting summaries. Fast, cheap and checked.
• What AI drafts and a person decides. Campaign concepts, social copy, customer messages, anything published under the brand name. AI proposes. A named person approves.
• What AI never touches. Positioning. Crisis communication. Anything that speaks for the brand in a moment of risk. Anything that needs you to know the room, the culture or the relationship. These stay with people, every time.

The lists will be different for every organisation. What matters is that they exist, that they are written down, and that the third list is respected even when a deadline says otherwise.

Where the people go

If AI takes the production, where do the people go? Up. Toward the brief, where the important decisions are made. Toward the customer, where the insight comes from. Toward taste, the ability to look at forty options and know which one is right, and why the other thirty-nine are not.

That shift is uncomfortable for teams built around making things. It is also the best news marketing has had in a long time. The part of the job that was always the most valuable is finally the part there is time for.

Where to start

• Write your three lists this month, and share them with every agency and freelancer you work with.
• Pick one workflow where AI saves the most time, and reinvest that time explicitly in better briefs.
• Review a month of your AI-assisted output side by side with your competitors'. If you cannot tell them apart, neither can your customers.

Bottom line

AI is the biggest production shift marketing has seen in a generation. Treat it as a production tool and you will make more of everything, including mistakes. Treat it as the reason to invest in judgement, and you will make fewer, better things that people actually remember.

So, which of your decisions would you never hand to a machine? Have you written them down?

Sources 3

1. The CMO Survey, Spring 2025, Duke University Fuqua School of Business.
2. Gartner, 2025 CMO Spend Survey, press release, 12 May 2025.
3. KPMG and the University of Melbourne, Trust, attitudes and use of artificial intelligence: A global study 2025.