AI SavvyFounders Standard Operating Procedure

Standard Operating Procedure

Generate on-brief images with ChatGPT commands

A shorthand system of prompt codes that turns a single line into product photography, lifestyle content, technical visuals, and ad creative without rewriting your prompt from scratch every time.

10 minute setup Beginner Last updated September 2026

Purpose

This SOP standardizes how the team requests images from ChatGPT. Instead of writing a fresh, lengthy prompt every time, each code below carries a full visual brief inside a single word. The result is faster output and images that stay visually consistent across a campaign or content batch.

Who this is for

Anyone on the team producing marketing visuals, product photography, or social content, including founders working solo who need fast creative without briefing a designer.

Tools needed

  • ChatGPT with image generation enabled
  • A clear description of the product or subject you are generating an image for
  • Brand reference details on hand: colors, tone, and any product specs worth including
Note
These codes are a personal shorthand convention built for this workflow. ChatGPT will not recognize the slash format on its own. You are teaching it the style through the description that follows the code, not the code itself.

Procedure

01

Identify your content goal

Before opening ChatGPT, decide what category of visual you actually need. This narrows down which code to reach for and stops you from writing a vague, catch-all prompt.

  • a. Product photography for a listing or website
  • b. Lifestyle or social content showing the product in use
  • c. A technical or explainer visual showing how something works
  • d. Advertising creative meant to sell or stop the scroll
02

Choose your base code

Open the full reference library below and pick the single code that matches your goal from step one. Treat this as your base layer. Everything else you add builds on top of it.

Example base code: /producthero

03

Add product and context details (optional)

A code alone gives the model a style to follow but nothing to render. Write one or two sentences describing the actual product, the setting, and any brand colors or materials that matter.

  • a. Name the product and its key material or finish
  • b. Describe the background or setting you want
  • c. Mention brand colors if they need to appear in the shot
Tip
Full example prompt: /producthero a matte terracotta ceramic mug on a warm cream marble surface, soft natural side light, minimal styling.
04

Stack a style modifier (optional)

If the base code needs a specific visual treatment, add one second code from a different category. Two codes is the reliable ceiling. See the combination examples for tested pairings.

05

Generate and refine

Run the prompt. Review the output against your brand guidelines before using it anywhere.

  • a. If lighting or color feels off, adjust the written description rather than rewriting the whole prompt
  • b. If the composition is close but not exact, ask ChatGPT to regenerate with one specific change
  • c. Save prompts that worked well so the team can reuse them

Full code reference

Codes are grouped by the outcome they produce. Use this as your lookup table during step two.

Product photography

CodeWhat it generates
/productheroPremium hero shot of the product
/productshotClean studio product photography
/detailshotMacro or detail photography on one feature
/flatlayTop-down arranged product composition
/packagingPackaging or product presentation shot
/3d-renderPhotorealistic 3D product render
/mockupProduct placed into a realistic mockup scene
/ecommerceMarketplace-ready product image

Lifestyle and social

CodeWhat it generates
/lifestyleProduct being used in a realistic setting
/ugcUGC-style authentic social image
/instagramInstagram-focused visual
/storyVertical Story or Reel creative
/socialpostSocial-media-ready creative
/testimonialCustomer or testimonial-style creative
/unboxingProduct unboxing scene
/thumbnailAttention-grabbing thumbnail

Technical and explainer

CodeWhat it generates
/explodedviewExploded visualization showing internal components
/cutawayCutaway view showing what is inside
/technicalTechnical or engineering-style visualization
/infographicProduct or feature infographic
/howitworksStep-by-step visual explaining the product
/featurefocusClose-up highlighting one product feature
/comparisonSide-by-side product comparison
/beforeafterBefore-and-after comparison
/problem-solutionVisual showing problem, product, then solution

Creative and advertising

CodeWhat it generates
/adscreativeHigh-converting ad creative or advertisement
/btsBehind-the-scenes, production-style image
/conceptCreative or conceptual advertising visual
/cinematicCinematic commercial-style image
/minimalMinimal premium product composition
/luxuryHigh-end luxury advertising aesthetic
/billboardBillboard or outdoor advertising creative

Combination system

Stack one base code from the product or technical tables with one style code from the lifestyle or creative tables. The base sets the subject. The second code sets the treatment.

/adscreative + /lifestyle
Lifestyle ad creative
/producthero + /luxury
Premium luxury hero shot
/explodedview + /technical
Technical exploded-view diagram
/ugc + /testimonial
Authentic customer-style testimonial
/productshot + /minimal
Clean e-commerce product photography
/bts + /cinematic
Cinematic behind-the-scenes shot
Warning
Stacking three or more codes in one prompt tends to produce a muddled result because the styles start competing with each other. Keep it to a base code plus one modifier.

FAQs

Codes work as style shorthand, not magic words. Pair the code with a specific product description, setting, and brand detail so there is enough for the model to act on.

Two is the reliable ceiling. Stacking three or more codes usually produces a muddled result because the styles compete with each other.

The underlying descriptions transfer to most image generation tools. The shorthand itself is a personal naming convention, so you may need to spell out the full description in tools that do not recognize custom shorthand.

Image models introduce natural variation between runs. Lock down lighting, angle, and background details in your written prompt if you need closer consistency across a batch.

Troubleshooting

IssueFix
Image looks flat or genericAdd specific lighting and material detail after the code. Vague nouns produce vague images.
Brand colors are missingName the exact color and where it should appear, for example on the packaging or the background.
Two codes produce a chaotic resultDrop back to one base code plus one modifier and rerun.
Product proportions look wrongDescribe the product's real dimensions or shape in plain language rather than relying on the code alone.
Output feels off-brandReference your brand guide directly in the prompt, including tone words like warm, minimal, or premium.

Want your prompt library built for you?

Get a done-for-you set of branded prompts inside the AI Challenge.

Join the AI Challenge →