Any good design, whether it be a photo, painting, or AI creation, needs an idea. With AI, that idea helps to form the basis of the prompt. Subject, location, action, all needs to be translated into a series of words that hopefully correlate into concepts the AI has been trained on.

While there is no perfect prompt, I’ve found stating the subject, action, and then location followed by keywords to describe the lighting setup, camera type and depth of field tend to produce good results. I start each prompt with an overall concept and see where the AI takes it. If it goes in the wrong direction, I’ll revise the prompt, expand on the concept and try try again.

Once the scene is set, the subjects looking good (and their hands looking mostly normal) there is still more work to do. Using the region tool in MJ to excess to change clothing, fix hands, or just add something to the scene. But that’s still not it. Additional edits are then made in Photoshop, using traditional Photoshop tools and new AI (more prompts!) to further enhance the image. In the case of ShopINSP, I also added in the product artwork. I finally take the image and run it through an AI enhancer to make the image look even better.