The democratisation of talent is no longer a promise: it is a fact. Today, the real bottleneck of creativity is rarely technical; it is almost always a question of direction. It no longer depends so much on the software you have in front of you, but on your ability to converse well with the machine, demand judgement from it and not accept the first answer as definitive.
AI does not work like an oracle. It is more like an extraordinarily fast assistant, capable of suggesting paths, but unable to decide for you which one is worthwhile. That is why working well with artificial intelligence is not about throwing out a brilliant prompt and expecting miracles. It is about iterating. Because if you don’t iterate, you’re not really creating: you’re just choosing from default options.
The first mistake when using artificial intelligence: obsessing over the perfect prompt
One of the most widespread myths is to believe that everything depends on writing the ideal prompt the first time around.
Experience shows the opposite: the first prompt matters, yes, but much less than it seems. The first response from any model is usually just a draft; sometimes useful, sometimes too obvious, sometimes surprisingly promising, but still immature.
The serious work begins afterwards: in the subsequent conversation, in the refinement, in the ability to ask for a second version that is better than the first.
A particularly effective trick to raise the bar is to ask the AI: ‘Rate this result from 1 to 10 according to the exact profile of my audience and tell me what it needs to reach a 9.’
That small change forces the model to stop being complacent and start behaving like a critical editor.
Over time, your personal library of proven prompts and structures ends up becoming one of your most valuable creative assets.
The best AI tools for creativity: don’t work with just one, learn to orchestrate them
Among the many artificial intelligence tools available today, few truly deserve to become a stable part of your creative flow.
The best results come when you stop treating AI as an isolated tool and start seeing it as a small, specialised team.
Each model has different strengths:
- ChatGPT remains a very solid reference for structuring ideas, summarising complex information, organising concepts, and turning scattered insights into clear content architectures.
- Gemini offers a particularly useful advantage in multimedia creativity: its integrated multimodality. It is one of the most natural options when you need an AI to interpret images, audio, screenshots, or visual materials within the same conversation.
- Claude stands out for its clean long-form writing, its ability to respect complex formats, and its more stable behaviour in lengthy documents.
The most effective practice is not usually to choose a single AI, but to combine several: one can help you create, another can detect weaknesses, and another can correct tone, clarity, or depth.
This small circuit often greatly improves the final result.
In a way, the next natural step is no longer to use AI, but to learn how to orchestrate them: deciding which will be your main AI and which will act as proofreaders or evaluators.
This will probably be one of the most important professional habits in the coming years.
But that deserves almost another article.
How to improve results with AI in professional environments
When AI enters the real world of work, the type of tools you need to master also changes.
Microsoft Copilot Pro is now a particularly recommended option in a professional environment, especially for those working within the Microsoft ecosystem.
Its integration reduces friction in documentation, email, synthesis and everyday operational tasks.
In Microsoft Excel, it is already particularly useful for logic, structure and formulas.
It also saves a lot of time in documents.
In Microsoft PowerPoint, however, we must be honest: no AI yet designs with true high-level visual criteria.
They help with structure and content, but the professional finish still depends on the human eye.
Generative image, music and video: when AI goes from idea to actual production
There is a lot of noise in visual creation, so it is important to distinguish between spectacularity and real usefulness.
Reve is one of those little-known tools that is particularly surprising for its realism, its treatment of light and its ability to move away from the artificial aesthetic that still appears in many visual generations.
At a more advanced level, there are tools such as Runway and Seedance
Here we are talking about creative workflows that are closer to demanding audiovisual production.
In music, the change has been just as profound.
Tools such as:
- Suno
- Udio
- MusicGPT
allow you to work on structures, atmospheres and variations in a matter of minutes, even in languages such as Basque.
In our own projects, such as Jordan Nights, a musical novel in development, AI has not changed the origin of the work, but it has greatly accelerated the exploration of new sound universes.
In video, the leap is also evident.
Flow with VEO is now a very powerful benchmark in visual control.
Sora is pushing cinematic language to a surprising level.
Grok offers an easier entry point for those who want to get started quickly.
But here the same rule applies again:
the first take is almost never the right one.
You have to ask for another light, another frame, less artifice, more narrative intention.
That’s where the real creative work begins.
Mistakes to avoid when working with generative AI: the stainless methodology
AI does not reward those who press first.
It rewards those who know how to persevere better.
Not settling for the first answer is already a creative discipline.
The first version rarely contains enough truth.
The second is better.
The third begins to really feel like yours.
Quick test: the five-minute test
Try a simple test today.
Take a text you wrote yesterday.
Run it through Claude or ChatGPT.
First, ask for harsh criticism: what is superfluous, what is wrong, what doesn’t work.
Then ask for two extreme rewrites:
- one for a ten-year-old child.
- another for a fifty-year-old executive.
Compare them.
Something very useful emerges: you begin to see which part of the message is structure, which part is tone, and which part is still truly your voice.
And that’s where the important part begins: when you stop asking for answers and start directing results.
AI can give you wings, but the direction is still human
Artificial intelligence removes many technical barriers.
But it does not invent a vision.
It does not decide what is worth keeping.
It does not distinguish for you between what has soul and what just sounds good.
The tool amplifies.
The voice remains human.
AI can give you wings. But leadership, fortunately, remains human.
