ChatGPT for Brand Designers: Fix Color Palette Drift in Product Images

You are a senior brand visual production specialist with 9 years of experience writing AI image generation prompts for consumer product companies where brand color accuracy in AI-generated product photography determines whether images can be used directly in marketing campaigns or require expensive post-production color correction. Help me write a product photography prompt so I can produce on-brand visuals and eliminate the color palette drift that currently makes AI-generated product images unusable without manual correction. My situation: - Product type and use case: [e.g., "a premium skincare serum in a frosted glass bottle with a brushed gold cap — primary use case is e-commerce product listing images and Instagram carousel posts"] - Brand color palette: [e.g., "the brand palette is warm cream #F5EFE6 (background), terracotta #C4704F (accent), dusty rose #E8C4B8 (secondary), and brushed gold #B8860B (hardware) — all four colors must be present and accurate in every generated image"] - Current drift problem: [e.g., "the background drifts from warm cream to cool white, the terracotta accent becomes orange-red, and the brushed gold cap renders as bright yellow or chrome silver depending on the lighting description used"] - Image tool and current prompts: [e.g., "ChatGPT image generation (DALL-E 3) — current prompts describe the color in generic terms like 'warm tones' and 'golden accents' rather than precise color language"] - Shot types needed: [e.g., "three shot types — hero product shot (centered bottle on textured cream surface), lifestyle context shot (bottle beside complementary props on a marble surface), and ingredient detail shot (close-up of serum texture with one key ingredient visible)"] - Lighting requirement: [e.g., "soft diffused natural light from the upper left, no harsh shadows, subtle gradient shadow beneath the product — the lighting must enhance the frosted glass texture without washing out the bottle's translucency"] - Output use: [e.g., "images will be used at 1:1 ratio for Instagram and at 4:5 for e-commerce listing — prompts must produce images that work cropped to both ratios without losing the key product elements"] Deliver: 1. A brand color language block — a precise color descriptor set that translates each of the four brand hex values into DALL-E 3 readable language, covering the specific warm or cool undertone, the surface finish (matte, satin, metallic), and the light interaction behavior (how the color responds to diffused light versus shadow) — designed to be appended to every product image prompt 2. A hero product shot prompt — a complete 150-word prompt for the centered bottle shot, incorporating the brand color language block, the frosted glass translucency description, the upper-left diffused lighting setup, and the textured cream surface specification, structured to minimize the probability of background color drift and cap color misinterpretation 3. A lifestyle context shot prompt — a complete 150-word prompt for the bottle-with-props composition, specifying the marble surface color temperature, the prop selection language (materials and colors that complement rather than compete with the brand palette), and the spatial arrangement that keeps the product as the clear visual anchor 4. An ingredient detail shot prompt — a 120-word prompt for the close-up serum texture shot, describing the serum color and viscosity, the lighting angle that reveals texture without specular glare, and the background depth-of-field treatment that keeps the brand palette readable in the blurred background 5. A negative prompt block for all three shot types — a list of eight exclusion terms preventing cool white backgrounds, chrome or yellow gold interpretation, harsh shadow, photographic noise, oversaturated skin tones, and prop colors that introduce non-brand palette elements into the frame 6. A color accuracy review process — a step-by-step method for comparing generated images against the brand palette using a free color picker tool, covering which five areas of each image to sample (background center, bottle body highlight, bottle body shadow, cap highlight, and prop surface), the acceptable hex deviation range for each area, and the decision threshold for accepting or regenerating 7. A prompt iteration log — a structured record for each generation attempt covering the prompt version used, the color accuracy score for each of the five sampled areas, the specific language change made to address the drift, and the generation outcome, building a prompt refinement history that prevents repeating failed approaches **Write every prompt component as a precise visual specification rather than a creative direction — product photography prompts fail when they describe mood or feeling rather than material, surface, light source, and color with the specificity that a commercial photographer would use in a shoot brief.**