Key Takeaways

The points worth keeping

• What really changed for Brazilian designers
• 1. Midjourney and image generation: ideation at scale
• 2. Stable Diffusion and ComfyUI: full control for those who need consistency

Professional prompt structure for Midjourney:

[MAIN SUBJECT] + [ACTION/COMPOSITION] + [VISUAL STYLE] + [ARTISTIC REFERENCE] + [TECHNIQUE] + [LIGHTING] + [COLOR AND MOOD] + [PARAMETERS]

Example — visual identity for a premium Brazilian coffee brand:
"Premium Brazilian coffee brand identity, minimal packaging design,
geometric patterns inspired by Amazonian indigenous art,
editorial photography style, matte textures,
warm terracotta and forest green palette,
sophisticated and earthy simultaneously,
flat lay composition on linen background --ar 4:5 --style raw --v 7"

Variations to explore:
--ar 16:9 (for digital applications)
--ar 1:1 (for social media)
--style raw (more realistic, less "AI")
--style scenic (more atmospheric)
--q 2 (higher quality, more time)

Practical workflow for presenting concepts:

1. Direction definition (you): before any generation, write three sentences about what the brand is, what it isn't, and the emotion it should convey.
2. Exploration generation (AI): 4 completely different visual directions with 4 generations each = 16 images in 15 minutes.
3. Curation (you): select 2-3 concepts most aligned with strategy.
4. Refinement (AI + you): deepen the selected directions, vary composition and palette.
5. Final execution (you): use the images as reference for work in Figma, Illustrator, or Photoshop.

Prompt for a Brazilian technology startup's visual identity:
"Brazilian fintech app UI concept, modern and trustworthy,
inspired by Brazilian modernist architecture (Oscar Niemeyer curves),
clean data visualization,
indigo and electric green palette,
professional but approachable,
UI mockup on smartphone --ar 9:16 --style raw --v 7"

Minimalist variation:
"Same concept but ultra-minimalist, white space dominant,
single bold color accent, Swiss design influence --ar 9:16 --v 7"

Expressive variation:
"Same brand but bold and expressive,
Rio de Janeiro street art influence,
vibrant colors, confident typography --ar 9:16 --v 7"

What not to ask Midjourney

• Final art: Midjourney generates inspiration, not high-fidelity execution. Using direct output as a logo or final piece is amateur hour—it's obvious.
• Text: Midjourney still struggles with text in images. Use it only as a placeholder.
• Character consistency: For recurring characters (mascots, brand illustrations), use Stable Diffusion with a LoRA specifically trained for your needs.

2. Stable Diffusion and ComfyUI: Full Control for Those Who Need Consistency

Best for: Brand mascots, recurring illustrations, proprietary style, product variations

While Midjourney is easier to use, Stable Diffusion via ComfyUI offers control that no simple interface generator can match. For studios that need style or character consistency, this is the choice.

Most-used workflows by professional designers:

Prompt for Stable Diffusion (A111 or ComfyUI format):

Positive:
"friendly Brazilian toucan mascot, flat vector illustration style,
bold outlines, limited color palette (5 colors max),
professional brand mascot, clean and scalable,
suitable for app icon and large format print,
Brazilian tropical colors, playful but sophisticated"

Negative:
"photorealistic, 3d render, complex textures, gradients,
too many colors, dark, scary, amateur, low quality,
blurry, deformed, extra limbs"

Settings:
Steps: 30-40 | CFG: 7-9 | Sampler: DPM++ 2M Karras

3. Adobe Firefly and Photoshop AI: Integration Into Your Existing Workflow

Best for: Designers already working within the Adobe ecosystem

For those who use Photoshop and Illustrator daily, Adobe Firefly integrated directly into these tools is the path of least resistance.

What actually works in Firefly in 2026:

• Generative Fill: Remove or replace elements in photos with professional quality. Removing a complex background object that would take 2 hours of manual selection now takes 3 minutes.
• Generative Expand: Extend photos to different formats without recreating—critical when you receive a landscape photo and need a portrait version for stories.
• Text to Vector (Illustrator): Generate editable vectors from descriptions. Quality is still inconsistent, but excellent for rapid prototyping of icons and ornaments.
• Background removal (Express): Automatic removal for cases where manual selection isn't worth the effort.

Effective prompt for Generative Fill in Photoshop:

To replace a background:
"Clean minimal studio background, warm off-white,
subtle gradient from top, professional product photography lighting,
no distractions"

To add an element:
"Brazilian tropical plants, monstera and banana leaves,
blurred background, natural bokeh, complementary to the main subject,
not competing with focal point"

To expand an image:
"Continue the scene naturally, match existing lighting and color palette,
seamless extension, same photographic style"

4. Figma and AI Plugins: Faster Product Design

Best for: Product designers (UI/UX), design systems, rapid prototyping

The AI plugin ecosystem in Figma has exploded. By 2026, the most adopted by Brazilian product designers are:

UI/UX workflow with integrated AI:

To create an app home screen with AI:

STEP 1 — Brief for Figma AI or Relume:
"Create a mobile app home screen for a Brazilian food delivery app.
Primary user: young professional, 25-35, São Paulo.
Key actions: browse restaurants, track order, access favorites.
Design language: clean and modern, warm food photography,
Brazilian colors (not cliché), accessible typography.
Include: header with location, search bar,
category filters (horizontal scroll),
featured restaurants (cards with image, name, rating, delivery time),
bottom navigation with 5 tabs."

STEP 2 — Review the generated structure
STEP 3 — Replace placeholder content with real content
STEP 4 — Apply the brand's design system
STEP 5 — Refine micro-interactions manually

What NOT to leave for AI in product design:

• UX decisions based on real user research
• Information hierarchy in complex flows
• Accessibility—always manually review contrast and sizes
• Micro-copy (labels, error messages, onboarding)—use AI for drafts, but have a human refine it

5. AI for Presentation Briefs and Copywriting

Best use: saving time on client communication, defending concepts with greater clarity

Designers spend far more time writing than they realize: briefing emails, concept presentations, design rationales, delivery reports. AI cuts this time by 60-70%.

Prompt for a creative design rationale:

"You are a senior art director presenting a creative concept to a client.

PROJECT: [project name]
CLIENT: [company and industry]
CONCEPT DEVELOPED: [describe the concept — e.g., 'visual identity inspired by Brazilian modernist architecture, with organic shapes, earth tones and forest green palette']
KEY CREATIVE DECISIONS: [e.g., geometric sans-serif typography, asymmetric grid, black and white photography with one accent color]

WRITE:
1. Introduction paragraph (maximum 80 words) — connects the concept to the brand's positioning
2. Rationale for each creative decision (2-3 lines per decision)
3. 'How this concept will translate into the brand's day-to-day operations' — 3 practical examples (e.g.: Instagram, business card, signage)
4. Anticipated response to the objection 'this feels too clean/simple' (in case the client questions it)

Tone: confident, clear, no design jargon the client won't understand."

Prompt for a creative brief based on a client conversation:

"Transform this briefing conversation into a structured document:

ORIGINAL CONVERSATION:
[paste the text or summary from the briefing meeting]

STRUCTURE INTO:
1. Brand context (what it is, who it's for, why it exists)
2. Project objective (what we want to achieve with this work)
3. Target audience (who will see this material — detailed persona)
4. Mandatory guidelines (logo, colors and fonts that must be respected)
5. Constraints (what must NOT appear — e.g.: competitors, elements the client hates)
6. Positive references (what the client showed and approved)
7. Negative references (what the client showed and rejected)
8. Expected deliverables (what we'll deliver and in what format)
9. Timeline and phases
10. Open questions (what I still need to ask the client)

Be specific. If information wasn't mentioned, mark it as '[TO CONFIRM]'."

6. Visual Reference Research with AI

Best use: discovering references you would never find on your own

One of the less obvious yet most powerful uses of AI for designers: requesting specific visual references.

"You are an art and design curator with deep knowledge of visual history.

PROJECT: [project description]
STYLE I WANT TO EXPLORE: [e.g., 'digital brutalism with influence from Brazilian graphic design of the 1970s']

SUGGEST:
1. 5 reference designers or studios (Brazilian and international) with links or searchable names
2. 3 historical design movements relevant to this project
3. 5 films or series with art direction close to what I'm looking for
4. 3 artworks or exhibitions that would influence this project
5. 5 English keywords to search on Pinterest, Behance and Are.na that will find what I want

For each reference: why it's relevant to this specific project."

"I want to create a visual identity for [brand] that is simultaneously contemporary and Brazilian — without clichés (no toucan, no green and yellow, no carnival).

Which designers, studios and Brazilian visual references show a sophisticated, non-obvious 'Brazilianness'?
Include: graphic design, architecture, contemporary art, fashion and photography.
Focus on work from the last 20 years that reinterprets Brazilian modernism with a contemporary eye."

7. What NOT to Leave to AI: Protecting Your Competitive Edge

This is the most important section. Designers who lose ground to AI are the ones who didn't understand this distinction.

What should never be automated

1. Creative strategy The question "which concept will move this audience to feel X and act Y" is strategy. AI can list options, but choosing which path to pursue — based on real user research, cultural knowledge and market intuition — is human.

2. Real user research AI can help analyze research data, synthesize interviews and identify patterns. But conducting the interview, reading body language, picking up on what the user didn't say — that's irreplaceable.

3. High-level art direction Knowing that a photograph with harsher side lighting will communicate more authority than one with soft diffused lighting for that specific client — this judgment comes from accumulated experience, not from prompt parameters.

4. Your personal signature style If you have a recognizable style, use AI to produce more work within that style — not to dilute it by trying to do everything. The designer with a strong personal signature style is becoming more valuable in a market flooded with generic output.

5. Ethics and cultural judgment AI reproduces past patterns. Recognizing when a concept reinforces stereotypes, appropriates culture problematically or excludes groups — this requires human awareness.

Career positioning for designers in the AI era

The design market in 2026 is splitting into distinct paths:

Skills that make designers most valuable in 2026:

1. Ability to brief AI precisely — writing prompts that yield professional results
2. Critical curation — knowing what works versus what's generic in AI output
3. Brand strategy — understanding the business behind the design
4. Process leadership — orchestrating clients, stakeholders, and tools (including AI)
5. Deep specialization — the more specific your niche, the less replaceable you are

Tools: the complete kit for designers using AI in 2026

Image generation

AI-assisted design

Productivity and process

Specific prompts for different project types

Visual identity

"You are a creative director specializing in Brazilian branding.

COMPANY: [name and description in 2 lines]
INDUSTRY: [industry]
TARGET AUDIENCE: [detailed description]
BRAND VALUES: [list 3-5 values]
COMPETITORS: [who they are and how they position themselves visually]

VISUAL POSITIONING PROPOSAL:
1. Brand archetype (which of Jung's 12 archetypes applies most and why)
2. Typography direction (category and characteristics — e.g., geometric sans-serif that conveys modernity without coldness)
3. Color direction (3-4 color palette with emotional and cultural justification for the Brazilian market)
4. Photography direction (style, lighting, composition, what should and shouldn't appear)
5. Overall feeling: 3 adjectives describing how the audience should feel when seeing this brand
6. One reference the client knows + what to take from it + what to avoid"

Communication campaign

"You are an advertising creative with experience in the Brazilian market.

PRODUCT/SERVICE: [describe]
CAMPAIGN OBJECTIVE: [brand awareness / launch / conversion / retention]
AUDIENCE: [detailed persona]
CORE MESSAGE: [what we want the audience to think, feel, or do]
MEDIA: [where it will run — social media, OOH, TV, digital]
TONE: [e.g., humor, emotion, rationality, urgency]

DEVELOP:
1. 3 completely different creative concepts (each with headline and rationale)
2. For the strongest concept: script for main piece (30s video OR static visual)
3. Adaptations for each indicated media format
4. Hashtag or campaign element for social media
5. The "hook" — what will make the audience stop scrolling"

Motion and video

"You are a motion design director specializing in content for Brazilian brands.

PIECE: [e.g., YouTube channel intro, product animation for Instagram]
DURATION: [seconds]
MESSAGE: [what to communicate]
STYLE: [e.g., dynamic typography, clean motion graphics, 2D illustration, mixed media]

MOTION SCRIPT:
1. Second-by-second (or scene-by-scene) breakdown
2. Visual elements in each scene
3. Transition type between scenes
4. Music/audio: what kind of soundtrack and why
5. Final call-to-action
6. Motion design references that approach this concept (searchable titles on Vimeo)"

Cases: Brazilian Studios and Freelancers Using AI with Results

Visual Identity Studio in Florianópolis

A 3-person studio that was handling 4-5 branding projects per month implemented AI in their ideation process in mid-2025. Result: they started closing 8-10 projects per month with the same team, without increasing overtime hours.

How their process works:

• Research phase: Claude analyzes the market and competitors, suggests directions
• Concept phase: Midjourney explores 4 visual directions in 3 hours (previously: 2 days)
• Presentation: Gamma AI builds the base presentation, team refines
• Execution: 100% manual in Figma and Illustrator — AI doesn't touch final execution

What didn't change: the brief, client listening, strategic choices, and final execution are entirely human.

UI/UX Freelancer in São Paulo

A senior product designer who worked as a contractor for startups added Figma AI and ChatGPT prompts to her wireframe workflow. She cut the wireframing phase from 3 days down to 1 day, taking on more projects per month. Result: 70% higher billing without increasing hours.

Her edge: uses AI for speed, but delivers what AI can't — real user research and UX decisions based on behavioral data.

Communication Agency in Recife

An 8-person agency implemented a process for generating campaign variations for retail clients. Before: 3-4 variations per campaign, requiring 2 days of work. With AI: 12-15 variations in 6 hours — more options for the client to A/B test, same team.

Your First Week Using AI as a Designer

Day 1 — Explore image generators Create an account on Midjourney or Leonardo.ai. Take an old project and generate 16 concept variations. No pressure to deliver — just exploration. Notice: how much of the creative direction came from you versus what emerged from the AI?

Day 2 — Use AI for an actual brief At your next briefing meeting, record the session (with client permission) and paste the summary into Claude. Ask it to structure the information into a briefing document. Compare the result with what you would have written manually.

Day 3 — Explore Figma AI Open a UI project and use Figma AI to batch-rename layers, suggest auto-layout settings, and search for components by description. Thirty minutes of exploration.

Day 4 — Write a rationale with AI Take a current project and ask Claude to write the rationale behind one of your creative decisions. Edit it to sound like you. Compare the AI version with your version. Which is better? Why?

Day 5 — Reference research For your next project, ask Claude for visual references before opening Pinterest. See how specific and useful the results are when you craft a well-structured prompt.

Days 6-7 — Reflection What did AI do that positively surprised you? What fell short of expectations? What's definitely better done without AI? Document your findings. That distinction becomes your competitive advantage.

Ethics, Authorship, and the AI Debate in Design

There's no way to wrap up this guide without addressing the elephant in the room.

The authorship debate: AI-generated images are created from billions of works by human artists, often without compensation or consent. This is a real ethical issue that the industry hasn't resolved yet. In Brazil, LGPD and the Marco Civil still lack specific regulations for generative AI, but that's starting to change.

What to do in practice:

• For commercial work: Adobe Firefly and Getty AI are the options with the clearest commercial licenses (trained on licensed content)
• For internal conceptualization: the debate is less urgent
• For final work delivered to clients: choose tools with explicit commercial use licenses

The replacement debate: Instead of pretending AI doesn't impact the creative job market, it's more useful to be honest — some junior designer tasks that were routine are now being absorbed by AI. That's reality. The smartest response isn't denial, but leveling up — in strategy, technical specialization, process leadership, or in mastering AI deeply.

Next Steps

The question isn't "Should I use AI?" anymore. The question is "Which parts of my process benefit from AI, and which require me to protect my human touch?"

Designers who answer this clearly and execute well on both fronts are the ones thriving.

Three actions for this week:

1. Identify 3 tasks in your current workflow that consume disproportionate time relative to creative output — these are candidates for AI.
2. Identify 2 areas of your work where your human judgment is the real differentiator — protect these from premature automation.
3. Sign up for a free account on Midjourney or Leonardo.ai and spend 1 hour exploring prompts on your next project.

For complete learning paths on AI for creatives, motion design with AI, and product design, explore the courses on TakeAICourse.com.

Create your free account and access exclusive content for designers and creative professionals.

Published on February 28, 2026 by TakeAICourse.com — The AI course platform for Brazilian professionals.