Artificial intelligence in the everyday work of a digital creator

In 2025 I built artificial intelligence tools permanently into my workflow, to automate content creation and data analysis. Ever since, the question I am asked most often is a variation on this one: so the machine does everything now?

No. And the long answer is more interesting than the short one.

What actually changes

Artificial intelligence has not made the work faster in the way people imagine. It has moved the time around. The hours I used to spend on mechanical tasks I now spend deciding, checking and connecting things. The total is not very different; the quality of what goes into it is.

A concrete example. Competitor analysis used to take days of manual collection. Today the collection and the first synthesis take hours, and I use the time saved to look at more competitors, compare more markets and form more hypotheses. I do not deliver sooner: I deliver a broader analysis.

Where I actually use it

Research and synthesis

Gathering and summarising information on the market, the sector and competitors, as a documented basis for strategy. This is where the gain is clearest, because it is about compressing a lot of material into little time. I always verify the sources and the conclusions: that is the part that cannot be delegated.

Content creation and copywriting

Assisted drafting and rapid exploration of message variants. The difference between a useful text and a generic one is not in the first draft, it is in the revision: tone of voice, factual accuracy, consistency with the client’s positioning. That part stays entirely human, and it stays the longest.

Data analysis and reporting

Reading larger data sets, spotting trends, preparing reports and analysing return on investment. Here the advantage is not speed: it is being able to look at more variables at once without losing the thread. But a trend is not an explanation, and no tool decides the difference between correlation and cause.

Workflow automation

All those repetitive tasks that require no judgement but consume attention. It is the least eye-catching use and perhaps the most valuable: it frees up mental energy for the decisions that matter.

The three limits I see every day

First: it produces plausible, not true. A generated text is convincing by construction. If you do not know the subject you have no way of telling a correct statement from an invented one delivered with the same confidence. Anyone without real expertise in the subject cannot use these tools safely: they can only hope.

Second: it tends towards the average. By the way it works, it returns what is most common. The result is averagely correct and averagely unremarkable. If you publish content indistinguishable from everyone else’s, you have not solved your visibility problem: you have moved it.

Third: it does not know your context. It does not know why that client turned down a proposal two years ago, what constraints their sector has, what works in their sales network. Context is precisely what you pay a professional for.

The principle I follow

Artificial intelligence does not replace expertise, strategy and judgement: it amplifies them.

Amplifying means the result depends on the starting point. If you have a method, AI makes it more efficient. If you do not, it lets you produce things that do not work more quickly. That is why I find the promise of fully automated marketing misleading: you automate the execution, not the thinking.

What you should ask whoever works for you

Not whether they use artificial intelligence — practically everyone will, and those who say they do not are often simply not declaring it. Ask where they use it, what they verify and who answers for the final result. If the answer is vague, the problem is not the tool.

In my case AI is one layer of the method, not the method. If you want to see how it fits into the four stages of a project, I have described it on the page about my way of working.