August 30, 2026AIExplore

How to Use Figma AI to Make and Edit Images

Generate and edit images inside Figma Design with prompts sized to the layout slot, then chain cleanup tools only when the crop or background needs help.

Figma AI

Figma AI can generate and edit images inside Design files from Actions or through the agent as part of a larger task. Write prompts that match the layout slot, exclude unwanted text or faces when needed, and watch shared AI credits. Pair generation with cleanup tools only after the subject is right. Overview: /explore/figma-ai.

This guide focuses on make and edit prompts sized to frames. For cleanup and vectorization see /blog/how-to-use-figma-ai-for-image-cleanup-and-vectorization. For agent chaining see /blog/how-to-use-figma-agent-for-conversational-design.

When make and edit images is the right job

Use image generation when the layout needs a unique fill and stock search is slower. Edit existing images when composition is close. Use cleanup tools when the problem is background, crop, or resolution rather than subject choice.

Step by step workflow

1. Measure the slot

Note frame size and whether the image is hero, card, avatar, or thumbnail. Prompt with that geometry in mind.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Slot card Frame: 1440 by 640 hero Role: background illustration Locks: no text, no faces. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Slot card Frame: 1440 by 640 hero Role: background illustration Locks: no text, no faces" that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Inputs:
Use the actual source material required for this task, together with the intended audience, destination, format, and quality requirements. Specify relevant files, text, URLs, structured data, reference assets, dimensions, language, tone, timing, model or generation settings, credentials, field mappings, or integration values when they apply. Keep secrets out of the example and use the product's supported credential or configuration mechanism.

Workflow:
1. Start with one representative input that contains the important characteristics of the real workload.
2. Configure figma-ai specifically for the requested task and make the important settings explicit.
3. Run the first pass and inspect the result against the objective.
4. Correct only the settings or source material responsible for a failed requirement instead of changing everything at once.
5. Validate the intermediate result before passing it to the next step.
6. When the result meets the requirements, save the successful configuration so the same process can be repeated consistently.
7. If this is a multi-step workflow, verify each handoff and keep a clear fallback or review path for failures.

Requirements:
The example must use supported functionality only. Do not invent product features, pricing, limits, integrations, model names, API behavior, or unavailable settings. Preserve important source data and formatting. Validate required fields before processing, handle empty or invalid input explicitly, prevent accidental duplicate processing where relevant, and stop or route the item for review when a required step fails. For credentials, use the supported secure configuration rather than placing secrets directly in the content.

Expected output:
The final result should directly satisfy the "Slot card Frame: 1440 by 640 hero Role: background illustration Locks: no text, no faces" requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

2. Generate or edit from Actions

Open Actions, choose make or edit images, write the prompt, generate, and place or replace the fill.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Prompt Abstract data visualization Palette: indigo and sand Style: minimal geometric Fit: wide hero. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Prompt Abstract data visualization Palette: indigo and sand Style: minimal geometric Fit: wide hero" that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Prompt Abstract data visualization Palette: indigo and sand Style: minimal geometric Fit: wide hero" requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

3. Inspect, then clean only if needed

Check crop and contrast on the real layout. Remove background, expand, or boost resolution only when the subject is already correct.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: QA [ ] Crop matches frame [ ] No accidental text glyphs [ ] Contrast works over UI [ ] Credits logged if batching. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "QA [ ] Crop matches frame [ ] No accidental text glyphs [ ] Contrast works over UI [ ] Credits logged if batching" that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "QA [ ] Crop matches frame [ ] No accidental text glyphs [ ] Contrast works over UI [ ] Credits logged if batching" requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Practical image prompts

Wide hero

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Wide hero illustration: abstract data visualization, indigo and sand, minimal geometry, no text, no faces, 1440 by 640.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Wide hero illustration: abstract data visualization, indigo and sand, minimal geometry, no text, no faces, 1440 by 640." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Wide hero illustration: abstract data visualization, indigo and sand, minimal geometry, no text, no faces, 1440 by 640." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Product lifestyle

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Soft lifestyle photo of a ceramic mug on a wooden desk, morning window light, shallow depth of field, no logos.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Soft lifestyle photo of a ceramic mug on a wooden desk, morning window light, shallow depth of field, no logos." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Soft lifestyle photo of a ceramic mug on a wooden desk, morning window light, shallow depth of field, no logos." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

App empty state

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Friendly empty state illustration of an open folder with a small plant, flat vector feel, transparent friendly, no text.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Friendly empty state illustration of an open folder with a small plant, flat vector feel, transparent friendly, no text." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Friendly empty state illustration of an open folder with a small plant, flat vector feel, transparent friendly, no text." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Avatar texture

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Soft abstract gradient avatar background, muted teal to cream, no faces, square crop.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Soft abstract gradient avatar background, muted teal to cream, no faces, square crop." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Soft abstract gradient avatar background, muted teal to cream, no faces, square crop." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Blog cover

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Editorial cover image of a quiet workspace, natural light, desaturated, space on the left for title later, no text in image.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Editorial cover image of a quiet workspace, natural light, desaturated, space on the left for title later, no text in image." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Editorial cover image of a quiet workspace, natural light, desaturated, space on the left for title later, no text in image." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Icon friendly mark

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Simple isometric server rack icon style illustration, two colors, centered, no background clutter.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Simple isometric server rack icon style illustration, two colors, centered, no background clutter." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Simple isometric server rack icon style illustration, two colors, centered, no background clutter." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Map mood

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Soft aerial city texture for a map header, dusk light, no readable street names.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Soft aerial city texture for a map header, dusk light, no readable street names." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Soft aerial city texture for a map header, dusk light, no readable street names." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Food card

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Overhead food photo of a grain bowl, bright window light, appetizing, square card crop.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Overhead food photo of a grain bowl, bright window light, appetizing, square card crop." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Overhead food photo of a grain bowl, bright window light, appetizing, square card crop." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Finance calm

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Calm abstract waves in navy and gray for a banking hero, trustworthy mood, no symbols that look like logos.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Calm abstract waves in navy and gray for a banking hero, trustworthy mood, no symbols that look like logos." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Calm abstract waves in navy and gray for a banking hero, trustworthy mood, no symbols that look like logos." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Kids education

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Gentle illustration of crayons and paper, warm light, playful, no characters from existing IP.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Gentle illustration of crayons and paper, warm light, playful, no characters from existing IP." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Gentle illustration of crayons and paper, warm light, playful, no characters from existing IP." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

More copyable examples

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Edit warmer Edit this image: warmer daylight, keep subject and crop, no new objects.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Edit warmer Edit this image: warmer daylight, keep subject and crop, no new objects." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Edit warmer Edit this image: warmer daylight, keep subject and crop, no new objects." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Edit remove clutter Edit this image: remove the stray cable on the desk, keep product and lighting.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Edit remove clutter Edit this image: remove the stray cable on the desk, keep product and lighting." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Edit remove clutter Edit this image: remove the stray cable on the desk, keep product and lighting." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Dark mode hero Dark mode abstract grid with soft glow nodes, deep charcoal base, no text.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Dark mode hero Dark mode abstract grid with soft glow nodes, deep charcoal base, no text." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Dark mode hero Dark mode abstract grid with soft glow nodes, deep charcoal base, no text." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Seasonal Quiet winter windowsill with mug and knit scarf, cool daylight, cozy, no sale badges.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Seasonal Quiet winter windowsill with mug and knit scarf, cool daylight, cozy, no sale badges." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Seasonal Quiet winter windowsill with mug and knit scarf, cool daylight, cozy, no sale badges." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Pattern fill Subtle seamless paper grain texture in brand sand, even lighting, tile friendly.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Pattern fill Subtle seamless paper grain texture in brand sand, even lighting, tile friendly." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Pattern fill Subtle seamless paper grain texture in brand sand, even lighting, tile friendly." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Device mock plate Soft gradient desk surface for a laptop mockup plate, neutral, no props.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Device mock plate Soft gradient desk surface for a laptop mockup plate, neutral, no props." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Device mock plate Soft gradient desk surface for a laptop mockup plate, neutral, no props." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Sports energy Dynamic abstract motion streaks in brand red, high energy, no athletes, no logos.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Sports energy Dynamic abstract motion streaks in brand red, high energy, no athletes, no logos." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Sports energy Dynamic abstract motion streaks in brand red, high energy, no athletes, no logos." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Scenario:
A realistic figma-ai workflow is being prepared for Use Figma AI for First Draft Explorations using figma-ai. The starting requirement is: Healthcare soft Soft abstract watercolor wash in sage and cream, calming, no medical symbols.. The work should be tested on a representative input first so the result can be reviewed before the same process is applied to the full project or production workload.

Objective:
Create a production-ready result for "Healthcare soft Soft abstract watercolor wash in sage and cream, calming, no medical symbols." that directly supports the article use case. The output should be specific, reviewable, and reproducible rather than a generic demonstration. Keep the requested goal as the primary outcome and avoid adding unrelated features, assumptions, or unsupported capabilities.

Expected output:
The final result should directly satisfy the "Healthcare soft Soft abstract watercolor wash in sage and cream, calming, no medical symbols." requirement and be understandable without guessing what was supplied or configured. A successful run should produce the intended output in the requested format, with the important values and workflow decisions clear enough to reproduce the result. If the workflow can fail, the expected behavior should also make the failure visible and provide a clear next action instead of silently producing an incomplete result.

Tips

Prompt to the frame size
Keep type in text layers
Exclude logos and faces when required
Clean up only after subject is right
Watch shared AI credits on batches
Verify contrast with real UI overlays

Verification notes

Image capabilities are documented in Use AI tools in Figma Design and agent feature tables for create and edit images. Credit usage follows the shared AI credits system. When plan access, seat rules, or credit behavior differ from memory, open help.figma.com and your AI balance. Limits change, so do not promise client timelines from outdated screenshots.

Keep a short session log beside the file: tool used, selection, prompt, and the one change you will try next. That habit turns four variants or parallel agent threads into learning instead of noise.

When Actions are missing, check whether an admin disabled Figma AI for the workspace before rewriting prompts. Access problems look like prompt problems until you confirm settings.

Treat the first AI pass as a sketch. Compare against your spacing grid, type styles, and accessibility targets before inviting stakeholders.

If you collaborate, store winning prompts on a shared page next to the library link. Reuse the skeleton and change only product nouns or channel constraints.

Separate exploration files from production libraries. Explore freely in drafts, then rebuild keepers with real components before handoff.

Close each session by naming frames clearly and discarding dead ends. A clean page saves time when someone asks for the pricing riff from Tuesday.

Prefer concrete product language over vague vibes alone. Screen type, primary action, and one constraint outperform a pile of adjectives.

Confirm AI credit balance and reset date before large image or agent batches. Shared credits power many Figma AI surfaces.

Match prompt aspect language to the frame. Wide heroes and square avatars fail for different reasons when the box is ignored.

Do not ask the image model to render long legal or marketing paragraphs. Keep type in text layers.

Chain cleanup tools only after the subject is right. Expanding a wrong image wastes credits.

Layout slot discipline beats poetic prompts. A perfect description still fails when the frame is a banner and the model returns a portrait.

Keep brand marks out of generated pixels whenever possible. Place logos as components so legal and brand updates stay editable.

Edit prompts should name what stays the same. Otherwise Figma AI may restyle the whole photo when you only needed a warmer grade.

Batch image generation late in the day only if you know your remaining seat credits. Shared credits also power text, agent, and other AI surfaces.

Contrast check with real type overlays before you fall in love with a fill. Soft midtones that look fine alone can kill headline readability.

When stakeholders ask for more magic, translate that into concrete controls: calmer palette, fewer objects, more negative space, or softer light.

Use Community or team photo search before generating if a real product shot already exists. Generation is for gaps, not for replacing approved photography pipelines.

After you pick a keeper, decide whether cleanup belongs in Figma or in an external retouch tool. Small fringing fixes are local. Complex compositing may not be.

Create a small prompt library per brand: hero abstract, empty state illustration, lifestyle still. Reuse skeletons and change only subject nouns.

Prefer fewer objects in generated scenes. Busy generations fight UI overlays and fail accessibility contrast more often.

When editing, change one photographic attribute per pass: light, clutter, or crop story. Multi attribute edits hide which instruction worked.

Export naming should include prompt intent. Future designers need to know why a fill exists before regenerating it.

Treat generated images as design ingredients, not finished brand photography. Final campaigns still need art direction, rights review, and often a human retouch pass.

If two generations are close, change only palette or object count on the next try. Full rewrites hide which lever improved the result.

Square social crops and wide web heroes should never share one uncropped master without a deliberate Expand or regenerate step.

Close image sessions by saving the winning prompt beside the frame as a comment. Regeneration without memory is expensive.

Common mistakes

Portrait prompts in wide banners
Long copy inside image prompts
Generating ten variants without changing one constraint
Expanding before the subject is approved
Ignoring credit balance on image heavy days
Shipping images with accidental glyph like shapes that look like text

Related Figma AI articles: /blog/how-to-use-figma-ai-for-image-cleanup-and-vectorization, /blog/how-to-use-figma-ai-for-first-draft-explorations, and /blog/how-to-use-figma-agent-for-conversational-design. Return to /explore/figma-ai for the broader guide set.