What Should You Create First in DrawNew? Start with the Problem

You have an idea, a photo, and a page full of AI tools. Where should you start? Choosing a model before deciding what needs to change can send you into unnecessary experiments.

Start with the material you already have and one useful outcome. The routes below connect common creative problems to a tool and a guide, so you can make a first attempt without treating every model as a separate project.

“I have an idea, but no image”

Use AI Image Generator to develop a new scene from text. Describe one subject, a setting, the light, and the intended layout. In the creator, select an available model and ratio, then review the credit estimate before generating.

If the result looks artificial, do not immediately add more adjectives. Read why AI photos look fake to diagnose materials, perspective, and shadows. If the result looks good but will not fit your page, fix the composition and ratio.

“I like the photo, except for one thing”

Use AI Image Editor, select editing mode, and upload the source. Separate the change you want from the details you want preserved. A cleaner background is a more focused first request than a new outfit, face angle, and lighting treatment together.

For a profile image, follow the headshot workflow. For a new color treatment, use the restyling guide. If the subject changes unexpectedly, the editing diagnostic helps identify the next correction.

“I need visuals for something I actually sell”

Start with a real product photograph and AI Product Photography. Decide whether the image is for a catalog, a banner, or a lifestyle scene. Those placements need different compositions.

The five product briefs show how to describe each job. The background replacement guide focuses on checking the outline, label, and shadows. For a menu item, follow the food photography workflow and compare the generated dish with the actual reference.

“I need a series, not one lucky image”

Save an approved reference and keep a stable brief for the elements that define the series. Change the scene-specific instruction separately. Use the recurring-character guide for a mascot or the social-image briefs for a coordinated campaign.

If you need to choose between models, use a small comparison with explicit acceptance criteria. The goal is to find a workable process, including cost and rework, rather than a favorite thumbnail.

“I have a still image that could become a short clip”

Open Image to Video, upload the approved image in the video editor, and describe the movement separately from the scene. Start with a slow camera move or restrained subject action. Select a supported duration and resolution and inspect the displayed estimate.

The Seedance 2.5 page includes reference images and motion briefs to explore. Its still images are inspiration and input references, not claimed video outputs. Check the entire generated clip for changes to faces, products, and background geometry.

What these guides are designed to do

Each article gives you a problem to diagnose, an actionable first attempt, and criteria for judging the result. Prompt examples are starting points, not promises of tested results or guaranteed accuracy. Where exact typography or original pixels matter, the guide explains when a separate editing workflow is more appropriate.

Browse all tools when you know the task, or the model library when you need to compare available settings. For more detailed help, explore tutorials, prompt briefs, and troubleshooting.