ChatGPT + DALL-E 3: how it works
On a ChatGPT Plus ($20/month) or Team plan, DALL-E 3 is built straight into the chat. You can write your prompt in plain English, and ChatGPT sits in the middle: it reads your brief and passes an optimised version to DALL-E. This is more than translation. ChatGPT's grasp of language turns a rough brief into a more accurate picture description.
With GPT-4o, image generation got better still. The model both reads and makes images, so you can upload an existing photo and say "recolour this product" or "add a person to this scene". The practical upside is that conversational briefing actually works. Write to Midjourney in loose, everyday language and the results drift; write to ChatGPT and it holds the context, then hands DALL-E a clean prompt. That gap shows most on cultural references, local aesthetics and images that need readable text. DALL-E is strong on text-in-image work like posters and packaging, on composition, and on cultural accuracy. It is weaker on fashion, editorial and high-contrast art direction.
Seven practical business use-cases
1. Social media visual: a weekly theme post, four variants in five minutes. Prompt: "Turkish tea, white marble surface, top-down shot, morning light, clean and simple, product-focused, no text."
2. Blog and newsletter header: custom imagery instead of generic stock. Prompt: "Cover image on technology and creativity, blue and purple tones, 16:9, for a corporate blog."
3. Product concept: a quick mock-up for the marketing deck of a product that doesn't exist yet. Prompt: "Minimalist perfume bottle, matte black packaging, white table, soft shadow, product photo, commercial use."
4. Packaging mock-up: a visual to show a client the packaging direction before anything goes to print. Prompt: "Coffee box packaging mock-up, kraft paper texture, minimalist typography, dark green accents, studio background."
5. Story template: a reusable 9:16 story frame. Prompt: "Story format, 9:16, dark background, logo area at the top, space for product image in the centre, modern and clean, brand colour yellow."
6. Deck cover slide: a visual for a company intro or sales document. Prompt: "Corporate deck cover, architectural photography style, city skyline, deep dark tones, professional mood."
7. Ad testing (A/B): ten thumbnail variants to find out which concept works before the real shoot. Producing variants conversationally in ChatGPT saves real time and budget here.
Prompt writing: tips that hold up
Give ChatGPT a full brief in your own words — something like "woman in her 30s, breakfast in a minimal Scandinavian kitchen, pastel palette, natural morning light, editorial lifestyle photography". ChatGPT optimises it for DALL-E. From there you can keep going by talking: "make this warmer in tone" or "keep the background simpler". That iteration loop is far more practical than Midjourney's.
A few things that help:
- Cultural accuracy: local references like "Turkish coffee", "tulip tea glass" or "iftar table" are recognised more reliably on DALL-E than on Midjourney.
- State the ratio: "1:1 square", "9:16 vertical", "16:9 horizontal" gets you output that fits the platform.
- Ask for variation: "give me three versions of this at different colour temperatures" pulls alternatives from a single prompt.
- Add a standard closing line: "no text, no watermark, commercial use" reliably pushes the prompt into image-generation mode.
- Use a style reference: "Architectural Digest style" or "Kinfolk magazine aesthetic" lifts output quality noticeably.
Copyright and commercial use: OpenAI's position
Under OpenAI's terms of use, the images you make with DALL-E 3 are cleared for commercial use and the copyright is yours from the moment you generate them. That differs from the limited commercial licence on Midjourney's free plan; on ChatGPT Plus or via the API, commercial rights are included by default.
It isn't zero risk, though. The source of DALL-E's training data has been challenged in court by some artists, and the same is true for Midjourney and Stable Diffusion. Adobe Firefly was trained only on Adobe Stock and licensed content, which puts it on firmer legal ground for enterprise and luxury-brand work. In practice the rule is simple: DALL-E is safe for internal decks and prototypes, while a campaign visual headed for publication is better served by Firefly or a studio shoot.
ChatGPT's limits: when to go professional
DALL-E 3 maxes out at 1792x1024 pixels — fine for screens, short of print quality. For large-format prints, billboards or a printed catalogue, ChatGPT is the wrong tool.
Character consistency is the bigger problem. Holding the same face, model or character across ten different scenes isn't reliable yet in DALL-E 3. If a campaign hero has to feature a brand ambassador or a recognisable face, a studio shoot becomes mandatory. The same goes for real product geometry — the exact proportions of a perfume bottle, the correct silhouette of a car, the detail on a piece of jewellery come out "close" rather than "correct". Shooting the product in the studio and then building the background and scene with Firefly or Gemini solves this.
ChatGPT is a good exploration and concept tool, not the spine of a campaign. It is excellent for showing a client which direction you're taking and not enough for producing the final visual.
How we work at PAM AI Studio
On a campaign we use ChatGPT at three points: sharpening the brief (text), the first 20-30 concept sketches (image), and internal deck prep and client sign-off on direction. At that stage its speed and easy briefing genuinely save time. Once the client approves the direction, the aesthetic gets worked up in Midjourney, and the final published visuals come out of a Firefly-and-studio combination. Each tool sits in its own layer. No black boxes.
To map out where your next campaign needs ChatGPT and where it needs a professional team, book a 30-minute discovery call with us.
Let's build it together.
We've been producing commercial film and photography since 2018, and we've been integrating AI into our workflow for three years. We mentor your team as we deliver: transparent process, documented decisions, no black boxes. Your AI production — set up for sustainable growth.
Email: [email protected]
Phone: +90 530 267 49 29
Studio: Yayıncılar Sok. 10/3, Seyrantepe · Istanbul