The spread of artificial intelligence in creative contexts is significantly reshaping the production of images, videos, and conceptual work.
Editorial
Prompt Design: the new ‘pencil’ in the creative process
Date
20 January 2026
Text-to-image and text-to-video systems allow visual content to be produced through a more rapid and open process, fostering the exploration of alternatives and heterogeneous visual languages.
In this context, the prompt acquires a genuinely design-oriented function: it requires specialised skills, evaluative capacities, and critical reasoning. Rather than replacing the Designer’s expertise, it becomes an integrated component of the workflow—one that must be understood, assessed, and managed with methodological rigour.
What prompt design is (and why it does not simply coincide with ‘good writing skills’)
Prompt design is the construction of instructions, constraints, and quality criteria aimed at guiding a generative system toward an outcome consistent with a creative objective. It is closely related to prompt engineering, yet it maintains a specific focus on design practice and visual language.
An effective prompt is not merely correctly formulated: it is built around an objective, a target audience, a cultural context, and an aesthetic direction.
The prompt can be understood in two ways:
- as an operational brief, oriented toward the immediate production of an output;
- as a design act, integrating intention, references, evaluative criteria, and responsibility.
To structure the work more clearly, it is useful to distinguish several fundamental components of a prompt.
- Intentions — They indicate the phase of the project in which the intervention occurs. A prompt may serve to generate initial ideas, develop alternatives to an already defined proposal, or refine and increase the coherence of a result.
- Levels of control — They define how detailed the prompt is. A high level establishes precise constraints; a medium level provides guided yet open indications; a low level leaves greater freedom to the system, which is useful in exploratory phases.
- Parameters — They concern the variables to be specified, such as style, tone, format, or technical and conceptual limits.
Distinguishing these aspects helps transform a generic request into a deliberate design choice
From brief to output: integrating prompt design into the design workflow
Prompt design is effective when it is embedded within a clear and repeatable workflow. An efficient sequence may be:
Objective → Data and reference gathering → Prompt writing → Generation → Selection of results → Revision → Archiving.
This is a cyclical process. Multiple versions are produced, alternatives are compared, and criteria are established to determine which direction to pursue. Even small changes in the wording of a prompt can generate markedly different outcomes, which is why keeping track of experiments is essential.
To maintain an orderly workflow, it is useful to:
- collect prompts in archives organised by project or typology;
- use clear naming conventions to distinguish different versions;
- annotate the tests performed and the results obtained;
- evaluate outputs according to defined criteria, such as coherence, originality, and adherence to the brief.
Prompt engineering for creatives: principles, patterns, and quality controls
In creative work, prompt engineering is effective when it makes the request clear and complete. A well‑crafted prompt defines the objective, makes constraints explicit, and organises information hierarchically, reducing ambiguity and misunderstandings. The use of targeted examples or references also helps orient the outcome.
To achieve a strong result, it is important to:
- assign the model a precise role, shaping its point of view;
- structure the request in successive steps;
- state explicitly what should be avoided;
- guide the style through coherent visual or cultural references.
A careful review is necessary after generation. Outputs should be evaluated in relation to their coherence with the initial idea, their stylistic quality, and their adherence to the brief, while paying attention to potential stereotypes or oversimplifications.
Generative systems are not always predictable and may exhibit biases or unexpected variations. Critical thinking therefore remains essential: the result must be selected, corrected, and integrated into the project, never accepted automatically
Applications in media design: text‑to‑image and text‑to‑video as tools for research and prototyping
**“Text‑to‑image and text‑to‑video technologies can be used in media design as support tools during the early stages of a project. They help build moodboards, visualise ideas, simulate scenes, and set up storyboards quickly and efficiently.
In editorial and graphic contexts, they make it possible to experiment with graphic systems, identities under development, and narrative sequences. They are particularly useful when multiple directions need to be explored before selecting the final one.
The process becomes faster and increases the number of available alternatives. This requires more careful control: producing many images does not equate to defining a solid proposal. Quality depends on the ability to select, organise, and synthesise results in a coherent way
Impact and responsibility: authorship, rights, bias, and the cultural sustainability of design practice
When discussing prompt design, it is essential to consider the impact and responsibility associated with what is produced. Authorship, for instance, becomes more nuanced: the person who signs the final result is not merely the one who writes the prompt, but the one who defines objectives, selects sources and references, sets constraints, and—above all—decides what to keep, what to discard, and how to refine the outcome.
In parallel, the issue of rights requires concrete attention: prompts, datasets, and outputs may incorporate elements traceable to existing works. Good practice therefore includes avoiding requests that directly imitate artists or brands, checking the licences of images and source materials, and documenting the process to increase traceability and protection.
Another critical point is bias: models can amplify stereotypes or flatten languages and imaginaries. More precise briefs, requests for alternatives, inclusivity criteria, and targeted human review help maintain quality, coherence, and contextual sensitivity.
Would you like to integrate Prompt Design into your educational path? Explore the IED courses in Visual Communication, Graphic Design, and Interactive Media, and develop solid design competencies for managing generative AI with methodological rigour and critical awareness.
In particular, the Master in User Interface Design and AI offers a structured programme that combines laboratory‑based teaching with thematic project work, adopting an approach grounded in practice and in the culture of design.
You may be interested in:
