Over the past twelve months, I have learned something simple: the more AI you use, the more creative direction you need. I learned this through very different real projects: Studio17’s relaunch, a complex editorial plan for a biotech company communicating R&D and regulations, a construction timelapse in Malcantone turned into a clear video narrative, and an institutional infographic for a Ticino municipality. AI entered the workshop and the pace changed. We cut time by over 60%*, but above all found a narrative coherence that had previously required weeks of refinement.

* Proprietary data 2024–2025; different contexts, limited samples.

The difference is not the “magic click” but direction: deciding what to ask the machine, where to let it run, when to stop it and how to give it room without losing voice and intention. This is what I call Augmented Creativity: AI not in your place, but beside you while you keep a firm hand on the wheel.

[1]
Emotional promptcraft
language you can touch

For premium brands, language is not just words: it is matter. When building prompts, I go beyond “minimal style” or “neutral palette” and bring in texture and light: porcelain glow, mid-century walnut, polarized sapphire. These sensory anchors guide the model towards a precise visual world, the right one for the brand.

It works like this: first I create a concise pre-brief with values, limits and off-brand metaphors to avoid. Then I request variations in separate steps: visual metaphor first, composition next, then adaptations for formats. Finally, I clean up with negative prompts: no clichés, no misleading ambiguities. AI accelerates, but we provide the meaning.

[2]
Assisted copywriting
human outline, fast draft, manual refinement

Writing with AI is not difficult; writing well still is. My workflow is straightforward: a manual outline covering objectives, tone and objections to address; a generative draft with legitimate references and no sensitive data in prompts; then human revision of style, consistency and micro-storytelling. Studio17 produced the result that struck me most: six short bullet-point pages became over 30 bilingual pages, with 12,000 Italian and English words in 11 days, and above all the same voice from the first paragraph to the last. The quantity is not what impresses; it is that, with AI in the right place, the narrative does not unravel.

[3]
Synthetic testing
testing small to make fewer mistakes

Before publishing, I prefer to “listen to the room”. I create synthetic personas and ask the model to assess two or three headline or hero variations using simple criteria: consistency, clarity and distinctiveness. It never replaces real testing, but helps remove noise. Then I check with 3–5 real people, such as colleagues or stakeholders, enough to see whether the chosen direction holds. It is an inexpensive filter that saves rework cycles.

[4]
Aesthetic impact metrics
the time the eye gives

Visual-heavy content needs to be measured with the same care used to design it. In GA4, I watch dwell time in rich sections, examine scrolling not only for depth but density – where people actually pause – and track quality clicks on galleries, details and soft CTAs. On Studio17, average time on page rose within a few weeks from around 0:48 to 2:05 (+161%). Not because of a special effect, but because the architecture, rhythm of images and tone of the writing spoke the same language.

If tracking tags or scripts are needed, the technical work is always handed over to IT. Here we remain focused on strategy and content.

[5]
HITL governance
more AI, more creative direction

When the scope grows – complex editorial plans, regulated content, multiple stakeholders – one rule applies: human-in-the-loop. A supervisor must verify sources, consistency and tone of voice, especially in scientific contexts. AI is an accelerator, not a guarantor of truth. By coordinating internal specialists, copywriters and language models, I can:

• cut drafting time by around −60%;
• maintain stylistic consistency across 50+ technical articles;
• produce coordinated infographics, including AI-checked palettes, then finalise them by hand.

In other words, AI amplifies the narrative and visual capability of those with solid creative direction. This applies to anyone aspiring to a premium position, not just established luxury brands. The advantage is not the “magic” of a button that solves everything, but the freedom to focus time and budget on strategy, concepts and the details that build a perception of quality.

These practices are more than productivity tricks. They are part of a process I am beginning to call Neogenerative Semiotics, in which AI is not just an execution tool but a co-participant in creating meaning. Every output becomes an input in a loop of continuous refinement.

________________

Want to optimise visuals and copy with AI while keeping the human touch? Let’s do a 30-minute audit.