Playbook
Generate on-brand images directly in Team-GPT chat
Generate on-brand images directly in Team-GPT chat
Upload any image for instant analysis, set your brand guidelines once, and create campaign visuals without switching tools or losing context
Step 1. Set up your Team-GPT Project knowledge
Navigate to your Project in Team-GPT (this could be for your company, a specific campaign, or a client you’re working with). Click the “Project knowledge” button to open the sidekick.
Add your essential visual context:
- Brand guidelines (colors, fonts, visual style rules)
- Logo files and usage guidelines
- Successful image examples from previous campaigns
- Target audience and brand personality
- Visual content pillars and themes
- Any upcoming campaigns or product launches
This context gets referenced automatically in every chat within the project, so you only need to set it up once.
Step 2. Upload your reference image to analyze
Find an image online that matches the style or concept you want to recreate. Save it to your device and upload it directly to your Team-GPT chat.
Ask the AI to analyze the image by typing something like: “Analyze this image and tell me about its key visual elements.”
Very cool website for visual inspiration and a good starting point I use is: same.energy.
You are an expert visual designer and image analyst with extensive experience in both traditional and digital art forms. Your specialty is deconstructing visual elements and providing detailed technical descriptions that allow other designers to replicate styles accurately, especially through modular, placeholder-based prompts.
Analyze the attached image and provide a comprehensive style breakdown that includes:
Start with a concise summary paragraph describing the overall impression of the image (style, tone, and subject). Then break down the image using the modular format below, ensuring clarity, realism, and design replicability:
-
Overall Style Classification
Categorize the image's style (e.g., minimalist, surrealist, vintage, editorial, cinematic, etc.) and justify why it fits that classification.
-
Composition Analysis
Evaluate framing techniques such as symmetry, rule of thirds, focal points, negative space, subject placement, and balance.
-
Color Palette
Identify primary and secondary colors, providing exact hex codes. Describe how color harmony or contrast contributes to the visual identity.
-
Texture and Pattern Characteristics
Describe surface textures (e.g., fabric, skin, background) and any repeating patterns. Note presence of grain, halation, noise, or analog-style imperfections.
-
Lighting and Shadow Techniques
Specify lighting direction, quality (soft, harsh, diffused), source type (natural, flash, ambient), and impact on shadow placement and realism.
-
Mood and Emotional Impression
Interpret the emotional tone, atmosphere, and narrative the image conveys. Identify how specific visual elements contribute to this mood.
-
Technical Specifications
If determinable, include resolution, aspect ratio, focus depth, camera angle, or lens effects. Highlight details that anchor the image in realism.
-
Medium Identification
Identify the creation medium (e.g., digital painting, studio photography, 3D render, mixed media) and note any signs that suggest it.
Format your response as a structured description that begins with a concise summary paragraph followed by detailed sections for each element above. Use the attached Playbook to better learn how to give this detailed response.
Conclude with a separate section titled "AI Generation Instructions" that translates your analysis into specific technical prompts optimized for AI image generation tools. These instructions should be detailed enough for a designer to recreate a similar style using AI tools. This should include:
- A fully assembled prompt using your modular structure
- Optional variations for styling, lighting, or mood for iterative testing
- Realism-focused guidance (e.g., "add visible pores and fine shadow gradation to avoid plastic look")
- Troubleshooting advice for common rendering issues (e.g., "if model looks too smooth, reintroduce surface texture or analog noise")
Example 1:
A seamless abstract wave pattern featuring flowing, organic shapes in main color: #047857 and secondary color: #052727. The design is modern, fluid, and minimal, with soft gradients and a high-resolution finish, perfect for digital and print branding.
Example 2:
Cinematic cyberpunk portrait of a man walking through a neon-lit futuristic city, nighttime scene, heavy rain, soaked trench coat, neon reflections on wet street, dramatic side profile, serious expression, vibrant pink and blue ambient lighting from neon signs, busy urban street with motion blur of rushing crowd, shallow depth of field, shot on 35mm film, detailed textures, moody atmosphere, Blade Runner aesthetic.
Optional additions for refinement:
- Camera settings: 35mm lens, aperture f/1.4 for shallow depth of field
- Lighting: ambient neon, rim lighting on edges
- Colors: neon cyan, magenta, purple, electric blue, hints of deep red
- Mood: introspective, tense, noir detective vibes
- Post-processing: analog film grain, slight chromatic aberration for realism
Example 3:
A crisp overhead flat lay photograph captured on a Leica Q2 with a 50mm prime lens, f/8 for full-frame sharpness and shadow detail. The scene is composed with high contrast and negative space, designed for graphic impact. Framed with the fries slightly spilling from a large branded white carton, caught mid-fall against a flat, bold scarlet red backdrop.
The branded white carton has "Team-GPT 3.0" written on it in a bold, sans-serif font, scarlet red, all caps, geometric and condensed, resembling 'Compacta Bold'.
The subject is a pile of golden fries, some airborne, arranged with dynamic tension as if frozen in time. Lighting is bright and direct, hard flash. Slight color grading for nostalgic warmth. Deep browns, acidic yellows, natural whites and sesame gold.
Post-processed with punchy saturation, vivid colours, ultra-crisp sharpness, and editorial style shadows. A touch of film grain. Styled like a still from an A24 foodie documentary, real, romantic, absolutely deliberate.
Example 4:
Cinematic racing scene, GoPro-style close-up shot of the rear wheels of a [ put ur car photo ] drifting around a corner, intense tire smoke, visible camber and wheel spin, asphalt detail with tire marks, motion blur and debris flying, aggressive drift angle, night setting with industrial lighting reflecting off bodywork, raw JDM drift energy
Example 5:
A man in stylish streetwear or fitted tuxedo jacket and sunglasses, moving through a packed nightclub. Focused expression, surrounded by sharply dressed crowd. Motion blur, long exposure light trails, yellow-orange streaks, gritty camera texture, raw chaotic energy looks.
Example 6:
A selfie of me (same face, sunglasses) with Kendrick Lamar at night in front of the glowing Petronas Twin Towers. Handheld, low-angle shot with motion blur, warm lighting, subtle grain, and a candid, cinematic vibe.
Example 7:
Night scene at a vintage 1980s American rest stop, glowing neon 'REST STOP' sign, wet asphalt reflecting lights, heavy fog in the background, parked retro car, dramatic lighting, cinematic atmosphere, film noir vibes, ultra-realistic, 35mm style
Example 8:
A hyper-realistic 4:3 editorial streetwear portrait of me with braided hair, wearing matte black Beats headphones, a diamond ring, and a clean tennis chain. I'm facing sideways, mouth open, pulling my lower lip to show off shiny gold-silver grillz. Lighting is soft and even. Background is plain gray for full focus. Bold, confident vibe in sharp
Example 9:
Surreal Y2K-style action shot of a young man mid-air in a dramatic leap, overhead fisheye view. He wears a vintage graphic tee, baggy jeans with a chain, and chunky sneakers. Windswept hair, stylized city blur below with retro cars and pedestrians. Bright daylight, high contrast, grainy like a 2001 skate game frame.
Example 10:
A 3:4 vertical studio photo of [subject] wearing [styling] in front of [background], shot with [lighting]. Emphasizes [mood/texture] with [optional: camera angle or motion].
Step 3. (Optional) Get a JSON profile for precise control
Here’s an advanced tip. For complete control over your image generation, follow up with this prompt:
“Give me a JSON profile of this image so I can generate another one in similar style and layout.“
The AI will create a comprehensive JSON profile with editable variables like:
- Color schemes and palettes
- Layout and composition elements
- Lighting and mood settings
- Object placement and sizing
- Typography and text elements
Step 4. Generate your brand-aligned image
Now ask the AI to generate an image using your brand guidelines: “Generate an image following this analysis but using our brand colors and guidelines from the project knowledge.”
The AI will create a new image that maintains the original’s composition while incorporating your brand elements automatically.
Create campaign visuals while campaign brainstorming
For campaign work, you can generate images directly within your brainstorming chats. Take your JSON profile from Step 3 and modify it with specific campaign elements:
“Use this profile to create an image with the headline ‘[Your Campaign Headline]’ and include our new product.”
The AI will generate campaign-ready visuals that include your text and branding, keeping your creative momentum flowing.
Images created in Team-GPT
Tips for better results
- Upload high-quality reference images: Better input images lead to more accurate analysis and recreation
- Be specific about brand elements: Mention exact colors, fonts, or visual styles you want emphasized
- Use the JSON method for consistency: Save successful JSON profiles for future campaigns to maintain visual consistency
- Test different variations: Ask for multiple versions with slight modifications to find the perfect fit
- Combine with campaign context: Generate images during strategy sessions so your team can see visuals alongside ideas
