How to Create a Mood Board with AI Textures and Typography in Figma

Last night, it took me a while to notice my coffee going cold on the desk in my San Telmo studio. A Figma canvas was open on screen, and a single line from the client's brief kept circling in my head: "Natural yet refined—but not too cold." This kind of request is the trickiest part of freelance life. When you exchange sensory direction through words alone, each side ends up picturing something different. So at some point I started building my mood boards around typography. I felt that the weight and texture of letterforms convey a brand's personality far more precisely than colors or photos.

Here I'm walking through that process—how to bring AI-generated textures and typographic experiments together inside Figma to create a mood board that presents a brand direction to a client.

Translating the Brief into the Language of Letterforms

When I receive the client's keywords, I don't jump straight to choosing fonts. First I think about what physical quality those words carry. Does "natural" feel like rough kraft paper, or like ink bleeding into water? Is "refined" the sharpness of metal, or the transparency of glass? Jotting down about five of these texture words in a notes app gives me the backbone for my prompts later when generating textures.

Next, I pick roughly three candidates from Google Fonts or my own type library. The important thing here is not to search for the perfect typeface. At the mood board stage you're showing a direction, not making a final decision. One Serif, one Sans-serif, and one Display face with a bit more character. I line all three up in a Figma frame and type out the client's brand name. That's the step where you see with your own eyes how the same word takes on a completely different temperature depending on the typeface.

Creating AI Textures

This is where AI image generation comes in. I mostly use Midjourney, and the key is to produce abstract textures that can work as backgrounds. I'm not making logos or illustrations—I'm building an atmospheric layer to sit behind the type.

The texture words I noted earlier become the prompt here. For example, on a recent branding project for a small wine bar in the Palermo neighborhood, I entered something like this: "worn leather texture, warm amber light, subtle grain, macro photography style." I pick two or three results with the right color tone and grain, then bring them into Figma.

I lower the opacity of the imported texture images or adjust the blend mode, then place them behind the type. My go-to combination is to drop the texture opacity to around 30–50% and layer the type on top in white or a dark tone. This gives an entirely different feel from type sitting on a plain white background. The letters feel as though they exist within a space.

Assembling Inside Figma

I keep the canvas layout simple. A single 1920×1080 frame is my basic unit. It's a size that's easy to take in at a glance when a client is viewing via screen share, and it's also convenient for exporting to PDF later.

The layout inside the frame is roughly as follows:

• Top half: a typography block with the brand name set at a large point size over the texture
• Bottom left: a comparison area for the three typeface candidates, showing the same phrase in each typeface
• Bottom right: four or five color palette chips, sampled from the texture with the eyedropper so the overall tone stays cohesive

One frame equals one direction. I usually create two or three directions for comparison, but I try not to go beyond three. The more options there are, the harder it becomes for the client to decide.

When combining type and texture, contrast is the thing I pay the most attention to. A clean sans-serif pairs well with a rough texture, while a handwritten-style face creates an interesting tension against a smooth texture. I've found that matching similar textures together often causes them to wash each other out.

Showing It to the Client

When sharing the finished frames, I don't send the Figma link right away. First I schedule a short video call and talk through it with the screen up. A mood board is the kind of deliverable that easily leads to misunderstandings if you just toss out images without explanation. You need to provide context—"I chose this typeface because it has this relationship with this texture"—so the client can properly evaluate the direction.

Many clients in Buenos Aires prefer meeting in person, so sometimes I grab my laptop and meet at a café over in Recoleta. When I show them the screen directly—resizing type in real time, adjusting texture opacity together—we reach that "Ah, this is the feeling" consensus much faster.

The whole thing usually takes about half a day. The cost for AI texture generation is just the Midjourney subscription, and everything else can be done even on Figma's free plan. This used to be a full-day affair—gathering reference images on Pinterest and collaging in Photoshop—but now I can spend that saved time exploring two or three more directions. And being able to swing by the neighborhood empanada shop with the hours I get back isn't bad either.