November 13, 2026AI’s role in the creative processItay SchiffCo-founder & Chief Creative Officer, Figma WeaveIllustration of a gold woven handbag decorated with blue flowers beside red and orange flowers with curling green stems, against a bright yellow background.

Which parts of the creative process can AI do for us—and which do we insist on keeping for ourselves?

Share AI’s role in the creative process

    Hero illustration by Jose Flores

    About a month ago, I signed up for a writing class. I wanted to crystallize my thoughts around AI in creative work by writing about them. The first assignment was simple: Write about a subject of your choosing. I started jotting things down, but very soon, I found myself asking ChatGPT to do the rest for me. My goal wasn’t to produce a polished piece, it was to articulate my own ideas. Yet I reached for AI within ten minutes—in a class I paid for!

    The experience highlighted something I think others are wrestling with, too. Many creative people today are using AI to do something they actually enjoyed doing, like writing, designing, editing, or directing. Now they’re working faster, but it doesn’t come without a deep feeling of conflict.

    Debates about AI and creativity often pit two camps against one another: the AI enthusiasts versus their opponents. But I think that split is reductive. I’m an AI enthusiast, a daily user building AI productsThree-panel product graphic demonstrating AI-powered style transfer. On the left, a "Transfer style" interface displays a target image of people carrying skateboards and a blurred portrait used as the style source. In the center, a workflow canvas labeled "Spring Campaign" shows a connected post-generation node producing a floral-themed social media graphic titled "Spring Arrivals." On the right, a gallery-style layout displays generated creative assets, including an illustrated handshake artwork and merchandise design cards, showing how transferred styles can be applied across a campaign.

    Connecting Figma and Weave

    Today, we’re announcing new ways your Weave creative workflows can live alongside your Figma frames.

    for a living. At the same time, I’m increasingly concerned about how easily I can surrender parts of the work that used to really matter to me. AI is changing the conditions of creative work, bringing enormous possibilities and real losses at the same time. To me, the question isn’t whether or not to use AI in our work, it’s about deciding which parts of the creative process we should automate, and which we insist on keeping for ourselves.

    Friction can be valuable

    Some processes in creative work can be horribly tedious. Think about a typical visual effects (VFX) shot: shooting on the green screen, removing trackers, building previs, adding 3D renders, pulling the green screen, refining the edges. That’s days and weeks of labor that doesn’t actually contribute to the artistic outcome. Most of us would gladly give up this kind of work. If we can automate these tasks with AI, we free ourselves up to work on bigger creative problems.

    Slide 1 of 61 of 6Man in front of green screen with monkey propThe many steps of the VFX processMan in front of green screen with monkey propMan in front of green screen with rendered monkeyMan in front of green screen with rendered monkeyMan on rendered background with rendered monkeyMan on rendered background with rendered monkey

    But some friction isn’t labor; it’s where the thinking happens. The time it takes to work through a problem can be time well spent. Maybe you’re editing a video. While you’re manually trimming a shot, you’re also considering its artistic impact. Or let’s say you are writing a creative brief. AI could handle it easily, but the process of drafting can give you clarity on your vision.

    Before we automate a step in our process, we need to ask ourselves what that step can do for us. Is it simply labor? Let the machine take it. Does it involve a thought process that could help us change our minds, take risks, or find meaning in the work we’re producing? Those are the parts we need to fight to protect. Workflow building is a new kind of craft.

    Generation isn’t authorship

    The problem is that the most important creative labor can be hidden in the tasks where it’s most tempting to use AI. Decision-making is hard. Not having the answer to a problem that needs solving is hard. Asking AI to do it for us is easy.

    Say you’re struggling with getting a frame just right. You can generate a new one in a few seconds, and have something that looks finished right away—the production quality seems high, and it takes care of the gap we were trying to fill. An AI model’s first output doesn’t look like the first drafts we used to get. The quality is deceiving, and it’s tempting to use as is. But anyone could have generated that same frame. The model generates material, but we still have to mold it. Otherwise, we all end up shipping the same mediocre outputs.

    We’ve been exploring this at Figma a lot recently. Take a campaign we just produced for Gaussian splat, a 3D scene rendering tool in Figma WeaveA “Spring Arrivals” campaign design shown in Figma Design beside its connected Figma node in Weave, with options to update the Figma design or review an update.

    The TL;DR on Figma Weave: Make your creative process repeatable

    Figma Weave brings AI generation and editing tools together into connected, reusable workflows, giving you a process your team can shape, refine, and repeat. Here’s what it unlocks for creative teams and how to get the most out of it.

    . We generated it using Seedance 2.5, a powerful video model that Weave hosts.

    The video features a creator exploring a 3D model as a metaphor for Gaussian splatting.

    While the generation technically could have happened in a single shot, the thinking around it took much longer. The campaign centers on a creator exploring a 3D model, which we used as a metaphor for Gaussian splatting. We started with a couple image inputs and a prompt. The v1 came out fast, and it was far from what we’d consider finished. So we kept riffing. We explored options for the story, the characters, and the styling. It was only then that Seedance could create shots we felt proud of.

    Keep the work that matters to you

    Building Figma Weave has helped me structure how I think of creative work today. On the Weave canvas, you can explore ideas openly, then turn your best ideas into reusable AI workflows. To me, this reflects the two modes creators operate in now. One is about exploring, producing first drafts, changing direction, and embracing uncertainty. I call this an “open mode.” But we also need a “closed mode” that’s more structured, where we commit to decisions, automate tasks, and make the work repeatable. Creators—and their tools—need to be able to switch back and forth. Workflows are most powerful when they preserve the choices a creator has made, while leaving room to reopen those choices when the work demands it.

    We can’t draw a universal boundary between what humans should keep and what machines should take on. A task that feels like menial labor to one person may be creatively useful to another. And that boundary may change from project to project. But before you hand something off to AI, examine it first. Be careful not to automate thinking away—especially when it’s most tempting.

    Itay Schiff is the co-founder and chief creative officer of Figma Weave. With 25 years in post-production and VFX, he leads Weave's vision of Artistic Intelligence—scaling ideation and media creation.

    XLinkedInInstagram

    Create and collaborate with Figma

    Get started for free