AI VISUALS / PRACTICAL PROMOTIONAL GUIDE 1 CREATOR FIELD GUIDE / SEPTEMBER 2026 A small team's guide to AI visuals Plan a clearer image, make a focused edit, and turn an idea into a short clip - without making your brief more complicated than the work. Promotional guide prepared for the operator of Dola AI, whisk-ai.co and nanobanana-ai.org. These are featured commercial websites, not independently ranked recommendations. This guide is not an academic paper or a hands-on performance test. Start with the post, not the tool A neighborhood cafe announcing its autumn menu does not need a whole imaginary campaign. It needs one image that feels right, enough empty space for a headline, and perhaps a short moving version for a social post. The useful starting point is a small creative brief: what should people notice, where will the visual appear, and what must stay accurate? Write those answers before uploading anything. A good brief might say: 'Show a ceramic mug on a quiet cafe table, with warm window light and clear space on the left for a headline.' That is a direction you can inspect. 'Make it amazing' is not. Three different starting points Use reference-led creation when you already know the visual direction. Use a text edit when one part of an existing image needs changing. Explore video when motion adds something the still image cannot explain. The next pages offer example briefs for each route; they are suggestions, not measured results. AI VISUALS / PRACTICAL PROMOTIONAL GUIDE 2 01 / SHAPE THE IMAGE Give each reference a job The studio at whisk-ai.co provides subject, scene and style reference inputs, plus text instructions. That structure can help turn a loose collection of inspiration into a more deliberate brief. Check the live interface for the available options and credit cost. Separate what you want to keep A subject reference answers what belongs in the picture. A scene reference suggests where it belongs. A style reference describes the overall visual treatment. Avoid giving three unrelated images equal importance and hoping the result will resolve the conflict. For the cafe example, start with a photo you own of the mug. If needed, add a table scene and a separate lighting reference that you have permission to use. Ask for a restrained composition rather than a collage of everything in the references. Example brief Use the mug as the main subject. Place it on a simple wooden cafe table near a window. Keep the handle visible. Use soft morning light and a neutral background. Leave clean space on the left for a short headline. Do not add lettering. Review before making more versions Compare the result with the source. Check the handle, rim, shadow and proportions. If you are representing a real product, an attractive invented detail is still an error. Change one instruction at a time so the next version answers a specific question. Visit the featured studio: www.whisk-ai.co AI VISUALS / PRACTICAL PROMOTIONAL GUIDE 3 02 / REFINE WHAT IS ALREADY THERE Ask for one clear change The editor at nanobanana-ai.org accepts image uploads and written editing instructions. For a small content task, a focused request is often easier to evaluate than a complete redesign. This guide does not establish any affiliation between the featured website and a model provider. Say what should not change Try: 'Replace the busy background with a plain warm-gray wall. Keep the mug's shape, handle and visible design unchanged. Match the existing light direction.' This makes the intended edit and the preservation requirements explicit. It does not guarantee that the model will follow every detail. Keep the original close by Look at the original and edited image side by side. Inspect small text, edges, reflections and any identifying marks. When those details matter, keep the original asset or finish the correction in an editor you can control directly. Add final headline text separately if generated lettering is unreliable. Use a simple approval rule Would a customer mistake the image for an exact photograph of something you sell? If so, verify every product detail before publishing. Do not present invented packaging, facilities or results as real. For people, use authorized photos and avoid edits that imply an endorsement they never gave. Try the featured image editor: www.nanobanana-ai.org AI VISUALS / PRACTICAL PROMOTIONAL GUIDE 4 03 / ADD MOTION WITH A PURPOSE Make the movement the brief Dola AI offers text-to-video and image-to-video modes. A practical way to explore the latter is to begin with an approved still image, then describe a small movement. Confirm the selected model, duration and credit cost in the live app before generating. Example motion brief A slow, subtle camera push toward the mug on the table. Gentle steam rises. Keep the mug and table stable. Preserve the quiet morning mood. No new objects, no lettering, no dramatic camera movement. Treat this as a concept to test, not a guaranteed output. Watch the whole clip, including its first and last frames. A convincing opening can still hide a changing handle or a distracting object later in the sequence. A short pre-publication check 1. Confirm you have permission to upload the source assets. 2. Check the current plan terms and commercial-use rights. 3. Inspect the output for changed products, faces and lettering. 4. Check the crop and leave room for platform interface overlays. 5. Add accurate captions and any necessary AI-content disclosure. 6. Save the approved source and final file so the choice is traceable. Free credits are not the same as unlimited free use or commercial permission. Dola AI currently distinguishes evaluation use from paid commercial licensing; confirm the applicable terms before using an output in an advertisement. No generation-speed or quality guarantee is made here. Explore the featured video tool: www.dolai.video Product descriptions checked against the three linked homepages on September 24, 2026. Features and terms can change. All example prompts in this guide are illustrative; no generated results or customer testimonials are claimed.