Nano Banana Universal Product Image
A universal prompt for Google's Nano Banana model that will combine any uploaded product photo with any image of a person.
Use This When
Creative production, ads, social content, mockups, visual testing.
Inputs Needed
Subject, style reference, composition, aspect ratio, brand colors, lighting, camera/lens, negative prompts.
Expected Output
Production-ready image prompt with subject, style, composition, lighting, camera, environment, aspect ratio, negative prompts.
The Workflow Prompt
You are a AI image prompt director and commercial art director.
Objective:
Nano Banana Universal Product Image
Context:
A universal prompt for Google's Nano Banana model that will combine any uploaded product photo with any image of a person.
Original task:
{
"mode": "image_to_image",
"input_images": [
{
"id": "person_image",
"description": "A person who will be shown interacting with the product."
},
{
"id": "product_image",
"description": "The product to be placed in the scene."
}
],
"prompt": "Create a photorealistic image where the person from 'person_image' is naturally interacting with the object from 'product_image'. The interaction should feel logical, comfortable, and realistic — for example, the person might be holding, using, wearing, or looking at the product. Ensure the lighting, perspective, and proportions of both images are harmonized so they appear in the same scene. Blend seamlessly with no visible artifacts from the source images. Maintain high resolution and an advertising-ready, polished look.",
"composition": {
"style": "photorealistic commercial photography",
"framing": "medium shot, product and person clearly visible",
"orientation": "upright, natural posture",
"background": "neutral or contextually appropriate setting that enhances the product"
},
"output": {
"resolution": "high",
"aspect_ratio": "4:5",
"quality": "advertising-grade"
}
}
Inputs I may provide:
Subject, style reference, composition, aspect ratio, brand colors, lighting, camera/lens, negative prompts.
Operating instructions:
- First, restate the objective in one clear sentence.
- If critical information is missing, ask up to 5 focused questions. If there is enough information to proceed, make practical assumptions and label them.
- Use a Exact Spec response style.
- Be specific to the business, audience, channel, and constraints provided.
- Avoid generic AI advice. Give concrete recommendations, examples, templates, copy, or steps I can use.
- When current facts, competitors, laws, prices, policies, or market claims matter, use current research and cite sources.
- Do not expose hidden chain-of-thought. Provide a concise rationale or decision summary instead.
- End with a short QA checklist that helps me verify the output.
Required output:
Production-ready image prompt with subject, style, composition, lighting, camera, environment, aspect ratio, negative prompts.
Caution:
Avoid over-polished AI visuals; specify real-world camera logic, imperfections, brand constraints, and negative prompts.
QA Follow-Up Checklist
After the AI returns its output, verify against:
- Output is specific to the provided business/context.
- Assumptions are clearly labeled.
- No unsupported claims without source checks.
- Next actions are clear and usable.
- Prompt includes camera/composition, motion, lighting, aspect ratio, and negative prompts.
Follow-Up Prompt
Now turn the result for 'Nano Banana Universal Product Image' into a client-ready version: tighten wording, remove fluff, add missing assumptions, and provide the next 3 actions.
Avoid / Cautions
Avoid over-polished AI visuals; specify real-world camera logic, imperfections, brand constraints, and negative prompts.
How Different Verticals Use This Workflow
Restaurant & Hospitality
A specialty coffee roaster wants ad creative showing a barista using their beans. They upload a barista headshot (with release) and a bag of beans. Output: 4 ad-ready composites of the barista pouring espresso with the bag visible on the bar, varied angles. Saves a $1,500 studio shoot. Constraint: brand colors on the bag are preserved exactly (critical for SKU recognition) and the barista's apron isn't a generic 'AI cafe apron'.
Retail & E-commerce
A men's grooming brand uploads a release-approved model photo and their beard oil bottle. Output: 6 ad-ready composites in bathroom and barbershop settings with the model applying the oil. Used as Meta ad creative variations to A/B test. Cost: $30 in API credits vs $4K for a campaign shoot. The brand explicitly locks the model's beard pattern (it's part of the brand) and the bottle label (regulatory).
Professional Services & B2B
A B2B SaaS company uploads a stock photo (with extended license for AI modification) and a screenshot of their dashboard on a laptop. Output: 4 composites of the model presenting the dashboard in a meeting setting. Used for trade show booth backdrops and case study covers. The dashboard UI is preserved exactly (not regenerated), which keeps the marketing accurate to the actual product.
Beauty & Personal Care
A skincare brand uploads a brand-approved model image and their serum bottle. Output: 5 composites of the model applying the serum at a vanity, warm side light. Used for the campaign launch on Instagram. The model's face isn't AI-distorted (a common Nano Banana failure with close-up faces), and the serum bottle label retains regulatory-compliant ingredient claims. Avoids a $6K beauty shoot for a small launch.
Local & Trade Services
A general contractor uploads a photo of his lead carpenter (with release) and a photo of a custom built-in he completed. Output: 4 composites of the carpenter installing the built-in, used in case study content. The wood grain and joinery on the built-in are preserved exactly (a contractor's reputation depends on craftsmanship being recognizable). Used in proposals and on the firm's site as 'process' imagery.
Frequently Asked
What does a great output for this look like specifically?
A composite where the person and product look like they were photographed together, not Photoshopped. Specifically: matched lighting direction on both, consistent depth of field, the same color temperature, and shadows on the product that match the room. If the model is lit from camera-left and the product casts shadow camera-right, the composite failed. Output should also preserve the product's actual branding and the model's identity (not 'AI's version of the person you uploaded'). Three test generations should produce three different scenes with consistent product geometry.
Is this safe to use with model release issues?
Only if you have rights to the person's image. Uploading a stock model headshot to composite with your product violates the stock license unless you've paid for the right modification tier. Uploading a real customer's photo without explicit written consent for AI manipulation is a legal problem in California (right of publicity) and the EU (GDPR). Use your own brand photoshoot models with AI-modification clauses in their release. Don't use celebrity photos under any circumstances — even for internal mood boards.
How is this different from product-images-from-uploaded-photo?
That prompt restages your product alone in a new context. Nano Banana adds a person interacting with it. Use that prompt for PDP shots, hero images, marble-and-light scenes. Use Nano Banana for lifestyle imagery, ad creative with human warmth, social content where 'someone using your product' converts better than 'product on surface'. They're complementary — many brands use product-images-from-uploaded-photo for PDPs and Nano Banana for ad creative.
What's the most common failure mode here?
The product comes out morphed. The model gets the person right but warps the bottle, mug, or device. Lock product geometry with 'preserve_shape_and_proportions: true' and add 'do not modify product' as a negative. Second failure: unnatural interactions — the person holding the product in a way no human would hold it. Specify the interaction explicitly ('person holding the bottle with two hands, looking at the label, not at camera'). Vague interactions produce uncanny-valley results.