How AI, design, motion and marketing are merging into one creative workflow

After finishing my studies, I realised something very quickly:

• I did not want to stop learning.
• If anything, finishing a course only made me more curious.

During my studies I learned how AI can support design, video, content creation and marketing. But once the assignments were finished, I started looking at what was happening outside the classroom.

I started reading articles from the Figma Blog about Figma Make, MCP and AI-assisted design-to-code workflows, including Laura Fehre’s “Workflow Lab: Deploying Designs Directly with Figma Make” and Riccardo Erra’s “Workflow Lab: Moving Between Design and Code with Agents.” I also followed Adobe’s Firefly updates, especially articles about agentic creative workflows and AI-assisted production, and watched practical tutorials such as TD Sunshine’s “Figma Make Tutorial & Best Practice Guide,” as well as Runway Academy’s Gen-4 courses on references, character consistency and generative video workflows. 

Figma AI moved from being simply a design tool in my mind to something closer to an agentic design environment: a place where an idea can evolve into interface concepts, editable layouts, prototypes and even code-backed experiences. Figma now connects design systems, AI agents, prototypes and development workflows through tools such as Figma Make and its MCP server. Figma

Notion AI became useful in a completely different way: as an AI-enabled knowledge workspace for research, project structure, notes, planning and content development. Instead of treating research and production as separate activities, information can increasingly stay connected to the project itself. Notion

Then there were platforms such as Higgsfield, with its AI-native image and video environment, cinematic workflows, camera-motion tools, character consistency systems and specialised studios; and Luma AI, whose current Ray models focus on generative video and motion. Higgsfield

Somewhere in that process, I noticed something bigger.

I started following product and design publications, official platform blogs and specialist tutorials: Figma’s own design and AI articles, Adobe’s Firefly and Creative Cloud updates, Runway Academy, and independent creators who break down practical workflows rather than only showing final results. 

And somewhere in that process, I noticed that the role of a creator is changing much faster than I expected.

Not because one tool suddenly does everything, but because more and more tools are starting to connect… 

From learning tools to understanding workflows

During my studies, a lot of the focus was naturally on learning separate skills: 

• How to create.
• How to edit.
• How to build a visual.
• How to use AI.
• How to think about marketing.
• How to tell a story.
• At first, I saw those things as separate disciplines.
• Now I see them more as parts of one creative system.
• A moodboard can become a campaign.
• A still image can become motion.
• A piece of music can become a visual identity.
• A prompt can become a prototype.
• A prototype can become a story.

And a story can become content for multiple platforms, that is where things became interesting for me.

The tools are changing

The development of creative software is no longer just about adding AI buttons to existing applications.

The applications themselves are changing.

Framer AI, for example, now combines an AI agent with responsive web design, editable layouts, CMS content, publishing and even custom code components. A prompt can become the beginning of a functioning website rather than simply a block of generated copy. Framer

Uizard focuses on AI-assisted UI/UX prototyping. Text can become editable multi-screen interfaces, while screenshots and hand-drawn wireframes can be transformed into digital mock-ups. uizard.io

Midjourney remains powerful for visual ideation, text-to-image generation, image prompting, style exploration and aesthetic direction. It is especially useful when the goal is not simply to produce one finished image, but to explore a visual world before committing to a direction. Midjourney

Adobe Firefly has expanded far beyond image generation. Its creative AI environment now covers images, video, audio, vectors, generative editing and moodboarding, while remaining connected to Adobe’s professional production ecosystem. Adobe

Canva AI and Magic Studio are moving Canva far beyond its original reputation as a template platform. AI-assisted design, image generation, video, text, brand tools, generative editing and editable layouts increasingly live inside the same Visual Suite. Canva

Recraft interests me particularly because it approaches AI from a graphic-design perspective. It combines raster generation with native vector generation, SVG output, mock-ups, editing and reusable styles. That makes it particularly relevant for brand systems, logos, icons and production assets where scalability matters. Recraft

And Leonardo AI combines generative imagery, image editing, motion and resolution enhancement in one creator-focused platform. Leonardo.ai

What surprised me was not simply what these tools could generate.

It was how much of the creative process they were beginning to touch.

At some point, switching between tools stops feeling like switching between unrelated applications.

It begins to feel like directing a creative pipeline.

A book changed the way I looked at it

Around the same time, I read Co-Intelligence by Ethan Mollick. Ethan Mollick — Co-Intelligence

What stayed with me was not the idea that AI should replace creative work.

Quite the opposite.

The interesting part is what happens when AI becomes something you work with:

• A collaborator.
• A sparring partner.
• A second pair of eyes.

A tool that can challenge an idea, offer an alternative, speed up exploration or help turn one concept into several directions, that also made me look differently at what I had learned during my studies, the technical knowledge was important.

But learning how to think with the tools may be even more important.

Prompt engineering is not just writing prompts

This is something I understand much better now than I did at the beginning, Prompt This is something I understand much better now than I did at the beginning.

The term prompt engineering can make the process sound as if everything depends on finding one perfect sentence.

Creative AI does not really work that way.

A serious workflow involves multimodal prompting, reference images, style references, composition, camera language, constraints, iteration, selection and evaluation.

It may involve prompt chaining, where the output of one stage becomes the input for another.

It may involve reference conditioning, where an image, layout or visual identity guides the next generation.

It may involve inpainting, where only a selected part of an image is regenerated, or outpainting, where the visual canvas is extended beyond its original boundaries.

And increasingly, it involves consistency control: keeping a character, product, colour system, environment or brand identity recognisable across multiple assets.

• You generate something.
• You compare it.
• You adjust it.
• You decide what works.
• You reject what does not.
• Then you move forward.
• That is not so different from traditional creative work.
• The difference is speed.
• And scale.