From surrealism to generative models
Generative process
Creative Gallery of Passion Fruit Cocktails
This gallery brings together fifteen variations on a passion fruit cocktail, created in an experiment with ChatGPT and Midjourney. Composition, light, materials and setting open up different readings of the same subject, changing the atmosphere, expectations and associations the image can evoke.
AI systems now allow textual instructions, visual references, alternative generation and targeted edits to be combined. The process develops through comparisons and revisions, guided by criteria for exploring and selecting proposals.
Each alternative may move closer to the communicative intention, introduce an interesting deviation or lead elsewhere. Making a choice requires considering the context, the people the image addresses and the interpretations it might encourage. Here, the gallery meets a central question of Neogenerative Semiotics: what conditions guide the formation of meaning, and how can its evolution be evaluated as it meets an audience?
AI for Branding
Analysis
Clear data, decisions without leaving things to chance. Structured research into markets, products and competitors defines positioning before any strategic choice.
- Market insights
- Clear gaps & opportunities
- Evidence-based decisions
Concept & Visuals
An identity that speaks at first glance. Names, taglines, visuals and mock-ups generated, tested and refined in rapid cycles.
- Ideation + A/B testing combined
- Style frames ready for stakeholders
- Consistency at every touchpoint
Editorial AI
A distinctive voice, scalable content. Editorial planning, SEO and multichannel storytelling created and localised in record time.
- Multichannel editorial plans
- Controlled tone in every language
- Strategic SEO & hashtags
Coding
From prompt to a prototype that compiles. AI writes and refactors front-end code, APIs or complex marketing scripts, accelerating the journey from demo to product.
- MVP in days, not months
- Refactoring without technical debt
- Automatically generated unit tests
Generation
Images, video, music: content on demand. From key visuals to sonic logos, I produce multimedia assets consistent with your brand voice in just a few prompts.
- Fresh content always available
- Reduced costs for shoots or stock assets
- Ad-ready formats
Strategies
A roadmap that becomes an engine for growth. Marketing plans, pricing and step-by-step guides generated and simulated with predictive models: less risk, more traction.
- Precisely allocated budgets
- Rapid scenario planning
- KPIs monitored in real time
Analysis
Clear data, decisions without leaving things to chance. Structured research into markets, products and competitors defines positioning before any strategic choice.
- Market insights
- Clear gaps & opportunities
- Evidence-based decisions
Concept & Visuals
An identity that speaks at first glance. Names, taglines, visuals and mock-ups generated, tested and refined in rapid cycles.
- Ideation + A/B testing combined
- Style frames ready for stakeholders
- Consistency at every touchpoint
Editorial AI
A distinctive voice, scalable content. Editorial planning, SEO and multichannel storytelling created and localised in record time.
- Multichannel editorial plans
- Controlled tone in every language
- Strategic SEO & hashtags
Coding
From prompt to a prototype that compiles. AI writes and refactors front-end code, APIs or complex marketing scripts, accelerating the journey from demo to product.
- MVP in days, not months
- Refactoring without technical debt
- Automatically generated unit tests
Generation
Images, video, music: content on demand. From key visuals to sonic logos, I produce multimedia assets consistent with your brand voice in just a few prompts.
- Fresh content always available
- Reduced costs for shoots or stock assets
- Ad-ready formats
Strategies
A roadmap that becomes an engine for growth. Marketing plans, pricing and step-by-step guides generated and simulated with predictive models: less risk, more traction.
- Precisely allocated budgets
- Rapid scenario planning
- KPIs monitored in real time
The video and all the icons on this page were co-designed with ChatGPT: I supplied the prompts, the AI generated the line art, and I then refined it and integrated it into the layout. A concrete example of how technology and creative direction can work together to produce original symbols in minutes, fully consistent with the visual style and context.
Generative process
From Static Rendering to an Overlay Video Loop
An evolution towards multimedia: a short looping video created from one of the gallery images. To produce this dynamic contribution, I transformed a static frame into a continuous animation using generative AI tools and motion design techniques. The original cocktail image was enriched with subtle movements and coherently animated details, such as a gentle sway of the liquid in the glass, changing light reflections on the drink’s surface and small particle effects that bring the scene to life. Once the raw video was generated, it was imported into Adobe Premiere Pro for initial editing, where I refined the sequence, adjusted its speed and ensured a smooth transition. The project then moved to Adobe After Effects for advanced post-production. At this stage, I applied special effects and professional corrections to enhance the visual quality: targeted lighting and contrast adjustments to emphasise the vibrant colours of the passion fruit and reflections on the glass; colour grading to align the palette with the rest of the page; and subtle glows or other effects, such as suspended light particles, to make the scene more realistic and evocative. The final result is a seamlessly integrated, striking video loop that emerges as an overlay from the original image. This process demonstrates how generative AI and specialist post-production can combine to create innovative, engaging multimedia content.
The page’s introductory video followed the same generative process, but with different prompts, lighting, settings and post-production, offering almost unlimited possibilities for generative variations and final results.
Governing AI processes
When an AI system plans activities, uses tools and carries out a sequence of actions, design also encompasses the conditions of its autonomy: objectives, available information, operational limits and review points.
Governing these processes means making responsibilities explicit and recognising when to continue, correct course or request a decision. The work must be open to observation, discussion and verification against the context and intentions that initiated it.
In my work, this concern meets Neogenerative Semiotics: the organisation of information, rules and handovers between people and systems also helps shape what is understood and what happens.
Annotated bibliography
For further exploration
Below is an annotated selection of books, papers and reports connecting art, semiotics, branding and Artificial Intelligence. These are the paths that have inspired – and continue to nourish – my working method. Enjoy reading and exploring!
Art & Representation
- René Magritte, La Trahison des images (1929) — the deception of the image as a starting point for discussing representation and meaning.
- Walter Benjamin, The Work of Art in the Age of Mechanical Reproduction (1936) — context, aura, reproduction: themes more alive than ever in synthetic generation.
- Italo Calvino, Lezioni americane / Six Memos for the Next Millennium (1985/1988) — lightness, exactitude, visibility: your “Precision + Warmth” has firm roots here.
- John Berger, Ways of Seeing (1972) — gaze, power and images: a valuable bridge between visual theory and brand storytelling.
Media & Communication
- Marshall McLuhan, Understanding Media (1964) — medium and message: a timeless compass.
- Rachel Lawes, Semiotics for Brands (2020) — applied semiotics, with concrete examples for reading brand signs and codes.
- Kristina Halvorson, The Content Strategy Toolkit (2015) — from editorial tactics to multichannel content governance.
- Roland Barthes, Mythologies (1957) — everyday “myths” as meaning-making machines for positioning work.
Branding & Design
- Alina Wheeler, Designing Brand Identity (5th ed., 2017) — a comprehensive handbook, from research to governance.
- Byron Sharp, How Brands Grow (2010) — empirical laws of growth, an antidote to received wisdom.
- Jenni Romaniuk, Building Distinctive Brand Assets (2018) — how to design, measure and protect distinctive assets.
- Jake Knapp, Sprint (2016) — a rapid process for testing brand and CX directions.
AI & Prompt Engineering
- David Foster, Generative Deep Learning (2nd ed., 2024) — current technical foundations (VAE, GAN, Diffusion, Transformer) for understanding what is “under the bonnet”.
- Jason Wei et al., “Chain‑of‑Thought Prompting Elicits Reasoning in LLMs” (NeurIPS 2022) — why and when explicit reasoning improves output.
- Schulhoff et al., “A Systematic Survey of Prompt Engineering Techniques” (2024) — an up-to-date taxonomy and best practices, rather than blog tips.
- Sahoo et al., “A Systematic Survey of Prompt Engineering in LLMs” (2024) — a comparative overview of techniques and applications.
- NIST AI 600‑1, AI RMF: Generative AI Profile (2024) — an operational reference for GenAI risks and controls, accompanying AI RMF 1.0.
AI × Branding
- Cillo & Rubera, “Generative AI in Innovation and Marketing Processes” (JAMS, 2024) — maps innovation stages and their impact on marketing.
- Grewal et al., “How Generative AI Is Shaping the Future of Marketing” (JAMS, 2024) — evidence and guidelines for adoption in marketing.
- MIT SMR, “When Generative AI Meets Product Development” (2024) — practical cases on ideation, concept testing and natural-language interfaces for complex tools. MIT Sloan Management Review
- HBR, “Turn Generative AI from an Existential Threat into a Competitive Advantage” (2024) — a strategic perspective on differentiation. Harvard Business Review
- Stanford HAI, AI Index 2025 — macro data on investment, adoption and productivity: context for brand decisions.
AI Governance & Policy
- NIST, AI RMF 1.0 (2023) — the Govern/Map/Measure/Manage framework for AI risks.
- NIST, AI 600‑1: GenAI Profile (2024) — actions and controls specific to GenAI.
- ISO/IEC 42001:2023 — the first AI management system (AIMS).
- EU Artificial Intelligence Act — Regulation (EU) 2024/1689 — the European regulatory framework, in force from 1 August 2024, with phased application.
- OECD AI Principles (updated 2024) — principles for trustworthy AI and regulatory interoperability.
by myself.