Como usar Nano Banana da maneira certa para criar imagens profissionais de IA

Many people start using Nano Banana by typing simple requests like “create a woman in a café” or “make a product photo.” The problem is that this kind of prompt usually generates generic images, with little personality and limited visual control.

Nano Banana, also known within the Google Gemini ecosystem, can go far beyond pretty images. It can create professional portraits, product photos, architectural visuals, landscapes, abstract art, and even edit existing photos with impressive precision.

But to get good results, there is one important rule: you need to guide the AI clearly.

The most effective structure presented in the transcript is based on six components:

subject, action, environment, art style, lighting, and details.

This simple formula helps turn vague prompts into complete instructions that produce much more professional results.

The Ideal Prompt Formula for Nano Banana

A common mistake is writing something too short, like:

“Woman in a café.”

This prompt may still generate an interesting image, but the AI has to guess almost everything: age, appearance, expression, lighting, photo style, environment, and scene details.

A better version would be:

“A young woman with freckles, smiling thoughtfully while holding a cup of coffee in a cozy café near the window, in a professional photography style, with natural window light, a blurred background, and a warm atmosphere.”

Notice the difference. Now the AI understands who appears in the image, what the person is doing, where the scene takes place, what visual style to follow, how the lighting should work, and which details make the scene feel more realistic.

This is the foundation for generating consistent images.

Professional Portraits

To create high-quality portraits, the secret is in the human and photographic details.

Instead of simply asking for “a businessman,” you can describe posture, expression, age, hair, eye contact, and lens type.

For example:

“Professional portrait of a confident executive with salt-and-pepper hair, looking directly at the camera, wearing a dark suit, photographed with an 85mm lens, blurred background, dramatic window lighting creating a rim light effect, and an editorial photography look.”

This kind of prompt helps Nano Banana create an image that looks like a professional photoshoot. Lighting, lens, and composition matter because they communicate depth, realism, and visual intention.

The main point is to avoid generic phrases like “good lighting.” It is better to explain the type of lighting you want, such as side lighting, soft light, window light, or dramatic lighting.

Product Photography

Nano Banana also works very well for creating commercial product images. This is useful for online stores, ads, social media posts, and brand presentations.

For product photos, the prompt should focus on sharpness, materials, controlled lighting, and composition.

A good example would be:

“Professional product photograph of a luxury watch on a dark surface, with controlled studio lighting, soft reflections, a macro lens to capture sharp details, elegant shadows, and a premium advertisement look.”

This type of prompt works because products need to be shown clearly. Unlike an artistic portrait, a product photo needs to highlight texture, finish, shape, shine, and details.

If you forget to request sharpness or a macro lens, the AI may create a beautiful image, but important details may appear blurry. For products, every word matters.

Architectural Visualization

In architectural images, it is not enough to simply describe the building. You also need to explain how the space, light, materials, and environment interact.

A more complete prompt could be:

“Modern minimalist house with large glass windows, surrounded by nature, photographed during golden hour, with warm light reflecting on the materials, wide composition, clean lines, and a sophisticated atmosphere.”

The time of day makes a big difference in this type of image. Morning light or sunset light can make the architecture look more valuable, with pleasant shadows and a more professional feeling.

For architecture, always think about three things: space, form, and lighting.

Realistic Landscapes

Landscapes require attention to depth, atmosphere, and composition. A good prompt should guide the AI through the foreground, background, and natural conditions.

For example:

“Alpine landscape at sunrise, with mountains in the background, a crystal-clear lake perfectly reflecting the sky, rocks in the foreground creating depth, soft golden light, and a cinematic atmosphere.”

Details like “crystal-clear lake,” “perfect reflections,” and “rocks in the foreground” help create a more visually organized image.

Without these elements, the landscape may look beautiful, but it can lack depth or focus.

Street Photography

Street photography needs to feel spontaneous. The challenge is to create an image that does not look too artificial or overly perfect.

A good prompt should use phrases like “spontaneous moment,” “natural expression,” “authentic urban scene,” and “realistic light.”

Example:

“Street photography showing a person walking on a busy sidewalk, natural expression, spontaneous moment, soft afternoon light passing through urban trees, background with real city context, and a documentary feeling.”

This type of image works better when the AI understands that the goal is not to create a studio pose, but to capture a realistic moment.

Abstract Art

Not everything needs to look like photography. Nano Banana can also create abstract images with impossible shapes, metallic textures, dramatic lighting, and artistic compositions.

In this case, the prompt should describe sensations, movement, and the relationship between shapes.

Example:

“Abstract art with flowing liquid metal forms, mixing organic and geometric elements, dramatic lighting creating depth, soft shadows, rich texture, and the appearance of a modern gallery artwork.”

The difference is that in abstract art, you are not describing a common object. You are giving the AI a creative direction to interpret.

Portrait Editing

One of the most useful parts of Nano Banana is editing existing images. You can improve a portrait without making the person look artificial.

The secret is to ask for subtle improvements.

Example:

“Enhance this portrait with natural skin smoothing, while preserving realistic pores and expression lines, slightly enhance the eyes, adjust the lighting to add more depth, and preserve the person’s natural appearance.”

The most important word here is “natural.” Many exaggerated edits make the image look fake. Good retouching should improve the photo without destroying the person’s identity.

Background Replacement

Background replacement is very useful, but it needs to be done carefully. If the background lighting does not match the lighting on the person, the result looks like a montage.

A good prompt should ask the AI to preserve the subject’s lighting and shadows.

Example:

“Replace the background with a modern and elegant office, while maintaining the natural lighting on the person’s face, preserving the original shadows, and ensuring clean edges around the subject.”

This helps the AI create a more convincing scene. It is not just about changing the background. It is about creating a new environment that matches the person.

Color Grading and Visual Mood

Color grading can completely transform the mood of an image. It changes the visual feeling, making the photo look more cinematic, warm, cold, dramatic, or vibrant.

Example:

“Apply cinematic color grading, with warm tones in the highlights, soft shadows, balanced contrast, and a professional look, while keeping the skin natural and avoiding exaggeration.”

This type of prompt works because it does not simply ask to “improve the colors.” It explains the kind of visual emotion you want.

Clothing Replacement

Nano Banana can also replace clothing in an image using a reference photo. But this feature requires very clear instructions to avoid artificial results.

Example:

“Apply the leather jacket from the reference image to the person in the main photo, maintaining natural body proportions, realistic fabric drape, compatible lighting, and coherent shadows.”

Clothing replacement works best when the new garment has a similar structure to the original clothing. Otherwise, the AI may create strange folds or unrealistic proportions.

Artistic Style Transfer

Another possibility is transforming an image into a painting, drawing, or specific artistic style.

Example:

“Transform this landscape into an impressionist painting, with visible brush strokes, soft color blending, an artistic atmosphere, and preservation of the main elements of the scene.”

The secret is balancing transformation and recognition. The image can change style, but it should still preserve its main identity.

Object Removal

To remove unwanted objects, it is not enough to say “remove this.” It is better to explain what should appear in its place.

Example:

“Remove the pole from the landscape and fill the space with sky and natural elements consistent with the rest of the scene.”

This guides the AI to think about the empty space after the object is removed. Object removal works better when you describe both what disappears and what replaces it.