JSON Context Profiles and Brand Case Studies

The AJCP Method (AI JSON Context Profile)

The AJCP method turns your creative direction into a reusable, shareable system. Instead of writing prompts from scratch each time, you define a JSON context profile that encodes your entire visual language.

The 2-Step Workflow

1. Define your context profile — a JSON object that captures scene, camera, lighting, color grading, mood, and imperfections.
2. Reference the profile in every prompt — paste or load the JSON as context before your specific shot description.

Example JSON Context Profile

{
  "project": "Autumn Editorial Campaign",
  "scene": {
    "setting": "abandoned European greenhouse",
    "time_of_day": "late afternoon",
    "season": "mid-autumn",
    "props": ["overgrown vines", "cracked glass panels", "weathered wooden tables"]
  },
  "camera": {
    "body": "Hasselblad X2D 100C",
    "lens": "80mm f/1.9",
    "film_simulation": "Kodak Portra 400",
    "aperture": "f/2.0",
    "focus": "subject eyes, shallow DOF"
  },
  "lighting": {
    "key": "natural window light, diffused through dirty glass",
    "fill": "bounce card camera-left for subtle shadow fill",
    "accent": "practical light from vintage Edison bulb in background",
    "quality": "soft, directional, warm"
  },
  "color_grading": {
    "palette": "muted earth tones with desaturated greens",
    "shadows": "cool blue-grey",
    "highlights": "warm amber",
    "contrast": "medium-low, lifted blacks",
    "saturation": "70% of natural"
  },
  "mood": {
    "emotional_tone": "nostalgic solitude",
    "energy": "quiet, contemplative",
    "references": ["Andrew Wyeth paintings", "Terrence Malick cinematography"]
  },
  "imperfections": {
    "film_grain": "subtle, organic",
    "lens_artifacts": "slight vignetting, minor chromatic aberration at edges",
    "skin": "visible pores, natural texture, no airbrushing",
    "environment": "dust particles in light beams, imperfect surfaces"
  }
}

Practical Workflow for Creative Directors

Follow these five steps to integrate context profiles into your production pipeline:

1. Mood board first — Collect 10-20 reference images that define the visual world. Extract the common elements into JSON fields.
2. Build the profile — Write your JSON context profile. Be specific about camera, lighting, and color. Vague profiles produce vague results.
3. Test with a hero shot — Generate your most important image first. If the profile doesn't nail the hero shot, refine the profile before moving on.
4. Batch with variations — Once the profile is dialed in, generate the full shot list. The profile ensures visual consistency across dozens of images.
5. Iterate the profile, not individual prompts — If something is off across multiple images, fix it in the profile. If it's off in one image, fix it in that image's specific prompt.

Jamey Gannon Workflows

Jamey Gannon, a leading AI creative director, has developed several workflows worth studying:

• Mood board first — Always start with a curated collection of reference images before touching any AI tool. The mood board IS the brief.
• Test individual SREFs — When working in Midjourney, test style references one at a time to understand what each contributes before combining.
• Layer personalization codes — Use Midjourney's --p codes in combination with SREFs to create layered, unique aesthetics that can't be easily replicated.
• Crop to remove unwanted elements — Rather than re-generating, crop the image to eliminate edge artifacts or unwanted compositional elements. Faster and preserves what works.
• Gemini as "Photoshop you can speak to" — Use Gemini's multi-turn editing for refinements that would traditionally require Photoshop: color adjustments, element removal, background changes, expression tweaks.

Brand Case Studies

Real brands are already using AI image generation at scale. Here's what the results look like:

Key Takeaways

• Brands using AI most effectively treat it as a production multiplier, not a creative replacement.
• The best results come from teams that combine strong art direction with AI generation — not from AI alone.
• A/B testing consistently shows AI-generated imagery performs at parity with or better than traditional photography when properly directed.

Exercise

Build Your Brand Profile

1. Choose a brand (real or fictional) and collect 10 reference images that define its visual identity.
2. Write a complete JSON context profile for that brand.
3. Generate 5 different shots using the same profile: a hero image, a product detail, an environmental shot, a lifestyle scene, and a social media crop.
4. Evaluate visual consistency across all 5. Would these feel like they came from the same campaign?
5. Refine your profile based on what you learned and regenerate.