# Create and check four business artwork concepts with ChatGPT Images

> Use ChatGPT Images to make four business artwork concepts from approved facts and assets, then reopen and check every saved file before deciding whether to use it.

By BIG CHANGE Editorial

Published: 2026-10-21T07:57:38.319Z
Updated: 2026-10-21T07:57:38.319Z
Canonical: https://bigchange.ai/blog/chatgpt-images-four-business-artwork-concepts-guide

![Conceptual charcoal illustration of a physical board holding four blank samples: an unlabeled chart, an ad card, a packaging reference sheet and a plain product box mockup.](https://bigchange.ai/api/media/file/chatgpt-images-four-concepts-board-hero-v2.png)
AI-generated conceptual illustration by BIG CHANGE.

ChatGPT Images can create and edit images from a written request or an uploaded reference. OpenAI's current [Images in ChatGPT guide](https://help.openai.com/en/articles/11084440-images-in-chatgpt) calls the experience **ChatGPT Images 2.5** and documents aspect-ratio, editing and Save controls. For a business, the useful result is four separately reviewed concepts: a factual infographic, an ad creative, a brand or packaging reference sheet, and a product or lifestyle mockup. Each needs its own approved inputs and checks.

This is a documentation-based procedure. BIG CHANGE did not use a ChatGPT account, upload business assets, generate an image or inspect a downloaded file for this guide. The checks below are for the reader to perform; the help page does not establish that any particular output will be accurate, rights-cleared or ready to publish.

## The big change

The current Images guide describes one interface for creating an image, revising it in conversation or with a selection, changing its aspect ratio and downloading it. A business can use those controls for four separate visual briefs, carrying each brief's source facts and rights into review. Reopen the saved file before its owner decides what happens next. OpenAI says a selected edit may change pixels outside the highlighted area, so every revision calls for another review of the whole image.

## Prepare the account and four source briefs

1. Use a ChatGPT account and workspace authorized for the material. OpenAI says ChatGPT Images is available on all tiers on web, iOS and Android. Image generation and uploads can have separate usage limits, and workspace settings may restrict what you can use. Check the controls and any limit notice in the actual account before planning four outputs. The optional **Images with thinking** mode has different plan access and is not required here. Templates are also optional and, according to the help page, unavailable in Work mode.
2. Before uploading, open **Settings → Data controls** and check the controls available in that account. In a personal account, turning off **Improve the model for everyone** keeps new conversations from model training, while saved chats can remain in history. If you submit feedback such as thumbs up or thumbs down, OpenAI says the associated conversation may still be used for training. Follow the organization's own rules for confidential assets, storage and retention. Do not move protected material to another account to work around a workspace restriction.
3. Ask the owner for a versioned input brief for each output. For the infographic, collect the exact claims, numbers, labels and their source records. For the ad, collect approved product facts, offer terms, call to action, destination and intended placement. For the brand or packaging sheet, collect the approved logo files, colors, type rules and permitted packaging words. For the mockup, collect the approved product photograph or exact description and an allowed setting. A link to a website is a lead to check, not proof that every product detail on a generated image is correct.
4. Make a private four-row worksheet before generation. In each row record the input revision and source, who owns or licenses every uploaded asset, permission to upload it, the intended aspect ratio, the prompt and references used, each edit, the downloaded file's real name and format, the result of reopening it, factual and rights findings, and the owner's decision. Leave the outcome fields blank until someone has checked the actual file.

OpenAI documents two starting routes: ask for an image in a conversation, or select **More → Images**. The Images sidebar also holds generated images. A request can describe the desired aspect ratio, or the reader can choose it with the aspect-ratio picker. Use only references the owner permits the workspace to process.

## 1. Make a factual infographic

Start a new image request with the approved fact table. State the audience, layout and exact labels. A reusable prompt structure is:

```text
Create one factual infographic from only the attached approved fact table, revision [ID]. Use these exact figures, units, dates, labels and source names: [owner-supplied fields]. Use [approved aspect ratio]. Leave out any claim or value that is absent from the table. This is a concept for human review.
```

Inspect the result at a size where the smallest text is readable. Compare every visible number, unit, date, claim, source credit and chart relationship with the approved table. If a value or visual comparison is wrong, select the image and describe the correction in conversation, or use **Select** in the editor to target an area. Recheck the entire image after the edit. If the source itself is incomplete, correct the source with its owner before generating again; do not ask the model to fill the gap.

Choose **Save** on the selected image to download it. Reopen the downloaded file in a local viewer and repeat the fact and legibility check on that file. Record the actual filename and extension. `01-infographic.png` is a useful name only if the downloaded file is really a PNG. A visually plausible graphic with an unresolved wrong figure is a rejected or unfinished concept.

## 2. Make an ad creative

Use a separate request and the approved offer brief. Specify the product, exact offer conditions, call to action, destination, permitted brand references and placement shape. For example:

```text
Create one ad creative concept for [approved product] using only the attached brief, revision [ID]. Include this exact offer and condition: [owner-supplied text]. Use this call to action and destination wording: [owner-supplied text]. Follow the approved brand references and [aspect ratio]. Do not add unapproved testimonials, endorsements, discounts or product claims.
```

Read every word and check the product depiction, offer conditions, logo and call to action against the brief. Check whether a person, customer quote or endorsement appears without authorization. Edit a supported element if needed, then inspect the whole image again, including the regions outside any selection. Save, reopen and record the actual downloaded `02-ad-creative` file and its true extension. The owner's decision concerns the concept; an ad launch or platform approval is a separate decision. For a related workflow using a different tool, see BIG CHANGE's [Pomelli campaign creative guide](https://bigchange.ai/blog/pomelli-checked-campaign-creatives-guide).

## 3. Make a brand or packaging reference sheet

Use the approved brand assets and exact permitted packaging copy. Request a sheet showing the approved mark, colors, type direction and package concept as references for discussion:

```text
Create one brand or packaging reference sheet from these owner-approved assets and rules, revision [ID]: [asset identities and rules]. Show only the permitted package words: [exact copy]. Use [aspect ratio]. Keep the sheet a concept for review; do not invent a seal, certification, claim or product variant.
```

Compare every visible mark, color, word, package label and product claim with the originals. An image model's rendering of a logo, color or font is not the approved source file. Revise any deviation through the documented edit controls and inspect the entire result again. Save and reopen the actual `03-brand-packaging-sheet` file, recording its true format. If the mark or required copy remains wrong, reject the concept. A raster reference sheet does not establish a vector identity, trademark clearance, approved label or print master. A reusable design system is a separate task, covered in BIG CHANGE's [Claude Design system guide](https://bigchange.ai/blog/claude-design-reviewed-brand-system-guide).

## 4. Make a product or lifestyle mockup

Use an approved product image or an exact approved description, plus a setting the owner permits. Ask for a concept that keeps the product's specified shape, label and color:

```text
Create one product or lifestyle mockup concept from the approved product reference, revision [ID]. Preserve these verified details: [owner-supplied details]. Place it in this permitted setting: [owner-supplied setting]. Use [aspect ratio]. Do not imply a real customer, event, use result or endorsement.
```

Inspect the product geometry, scale, label, color and background. Check whether the image could be mistaken for a photograph of an actual product use or a real person's endorsement. Edit a specific problem if the controls allow it, then inspect the whole image again. Save, reopen and record the real `04-product-lifestyle-mockup` file and extension. Mark it as synthetic or illustrative wherever its later use could otherwise imply documentary evidence. Reject the file if the product or scene remains misleading.

## Finish with four file decisions

Put the four downloaded files beside the four worksheet rows. Open each file from its saved location, check that it serves its assigned role, and compare its contents with that row's source revision and rights record. For each, the owner records **accept as a concept**, **revise** or **reject**, with the remaining issue. An accepted concept is not an automatic approval for advertising, product listings, brand adoption, printing or public release.

If Images or an edit control is unavailable, check the account and workspace access shown at that time. If a generation limit appears, record the exact notice and resume when the account permits it or after an owner-approved plan decision. If Save seems uncertain, inspect the Images library and the download location before generating a replacement. Preserve the downloaded format; changing an extension does not convert a file. If any of the four checked files is missing, the four-output task is incomplete.

OpenAI's help page states tier availability but gives no universal four-image allowance or per-image price for this procedure. The Free tier FAQ says image creation and uploads have separate limits and ChatGPT displays notices when a limit is reached. Check current account terms, plan charges and limits for the actual account. OpenAI's applicable [individual terms](https://openai.com/policies/terms-of-use/) or [business agreement](https://openai.com/policies/services-agreement/) require rights to supplied inputs and review of output. Their allocation of output rights between the account holder and OpenAI does not clear someone else's logo, photograph, font, likeness or product claim. OpenAI also warns that output can be inaccurate or similar to another user's output. Human review is necessary before sharing or using any concept.

## Sources & further reading

- [OpenAI, “Images in ChatGPT”](https://help.openai.com/en/articles/11084440-images-in-chatgpt). The current help page, checked 21 October 2026, names Images 2.5 and documents entry points, uploads, aspect ratios, editing, Save and Images storage. It warns that a selected edit can affect other areas. It does not identify a backend model ID or promise accurate business artwork.
- [OpenAI, “Data controls in ChatGPT”](https://help.openai.com/en/articles/7730893-data-controls-in-chatgpt). Explains account-dependent controls, the setting for new conversations, the feedback exception and the continued presence of saved chats.
- [OpenAI, “ChatGPT Free Tier FAQ”](https://help.openai.com/en/articles/9275245-chatgpt-free-tier-faq). Explains separate image and upload limits and in-product limit notices; it does not supply a universal allowance for this four-file task.
- [OpenAI, Terms of Use](https://openai.com/policies/terms-of-use/) and [OpenAI Services Agreement](https://openai.com/policies/services-agreement/). The applicable account agreement governs input permission, output rights between the parties and output review. Neither is third-party clearance for commercial artwork.
- [OpenAI, Usage policies](https://openai.com/policies/usage-policies/). Relevant to misleading likenesses, impersonation and deceptive concepts.

## Sources

- [Images in ChatGPT](https://help.openai.com/en/articles/11084440-images-in-chatgpt) — Primary source for current documented image-generation/editing and Save controls. Access, plan limits and workspace settings must be checked for the reader account.
- [Data controls in ChatGPT](https://help.openai.com/en/articles/7730893-data-controls-in-chatgpt) — Account-dependent data controls; personal model-improvement toggle applies to new conversations and does not delete saved chats. Feedback such as thumbs up or down may allow training on the entire associated conversation after opt-out.
- [ChatGPT Free Tier FAQ](https://help.openai.com/en/articles/9275245-chatgpt-free-tier-faq) — Separate image-generation and upload limits and in-product notices; no fixed four-image allowance.
- [Terms of Use](https://openai.com/policies/terms-of-use/) — Individual account input permission, output rights between parties, potential similarity and human review duty.
- [OpenAI Services Agreement](https://openai.com/policies/services-agreement/) — Business customer input permission and output review duty; account/order terms govern fees.
- [Usage policies](https://openai.com/policies/usage-policies/) — Relevant prohibitions on deceit, impersonation and confusing likeness use without consent.