Visual Brief Template for Image Production

Why you need a brief before writing an image prompt

A two-word prompt such as "perfume photo" sometimes gives you a beautiful image, but it is the tool's image, not yours: colours you did not choose, an angle that does not fit the post, no room for text, and a style that changes every time. A visual brief settles these decisions before generation, so the image prompt becomes a simple translation of the brief, and review becomes a comparison against written criteria rather than shifting taste.

A brief is also useful when you are not generating the images yourself: you hand it to a designer or a colleague, or come back to it a month later to produce a new image in the same series.

The working rule: a consistent style matters more than a beautiful image. A good image that looks like the rest of your images serves your brand better than a stunning one that looks like nothing else you publish.

From brief to prompt: the anatomy of an image prompt

Any good image prompt can be broken into eight parts, and the brief is built on them:

[main subject] + [action or pose] + [environment] + [style] + [lighting] + [angle and lens] + [colours] + [aspect ratio]

The brief adds what does not go directly into the prompt but governs the work: purpose and use, materials and texture, references, text placement, exclusions and review criteria.

The brief, field by field

1. Purpose and use

Where will the image appear, and why? A product-page hero, an Instagram post, a story cover, a printed banner. The use sets the aspect ratio, the resolution and the space for text; the purpose sets what must catch the eye first.

2. Subject

What is the main object or scene? Be specific about shape, size and the details that matter. "A loaf of bread" is generic; "a round whole-wheat loaf with a light coating of za'atar and sesame, a quarter cut away to show the crumb" is specific. If the product is real, write down the details that must not change: shape, colour, label, number of pieces.

3. Action or pose, and environment

What is happening, and where? Is the product sitting on a surface, held in a hand, or in use? What is the setting: a kitchen, a wooden table, a shop shelf, a plain colour background? The environment carries meaning: a wooden table suggests home, a flat colour background suggests something modern.

4. Composition

Where does the subject sit in the frame, and how much of it does it fill? Centre or on a third? Close-up, medium or wide? Are there foreground elements that add depth? Also state what must stay empty, which links directly to text placement.

5. Camera: angle, lens and depth

Even for generated images, camera language gives the model signals it understands: angle (eye level, directly overhead, low angle, 45 degrees), lens (wide to show the space, 50 mm close to natural vision, 85 to 100 mm for products and faces with less distortion) and depth of field (shallow to blur the background and isolate the subject, or deep to keep everything sharp).

6. Lighting

Light sets the mood more than any other element. Specify the source (side window, morning sun, soft studio light), the direction (side, back, front) and the character (soft, hard-shadowed, warm, cool). For example: "warm morning light from a window on the left, soft shadows, a slight sheen on the surface".

7. Materials and texture

What are the things in the image made of, and how should they feel? Raw olive wood, white marble, crumpled linen, clear glass, recycled paper. Materials are what make an image feel "real" or "plastic". Also list excluded materials if they clash with your identity, such as very glossy surfaces for a brand that wants a natural feel.

8. Brand colours

Write the colours by name and code if you have them, and give them roles: the dominant colour, the secondary colour and the accent that appears in a small area. Models do not always respect codes precisely, so describe the colour too ("dark olive green") and check the result by eye or with the colour picker in your design software.

9. Style and the fixed style sentence

Is the image realistic photography, a flat illustration, 3D, watercolour? Write the style as one fixed sentence that you paste at the end of every prompt for the same brand. This "style sentence" is the single most important tool for consistency across images.

10. References

Reference images that show what you want: earlier brand images, a real photo of the product, a colour palette. For each reference, state exactly what we take from it (only the lighting, only the composition). Do not use other photographers' or brands' work as something to copy, and do not ask for the style of a living artist or a known brand.

11. Text placement

If the image will carry a headline or logo, say where, and ask for calm empty space there. In Arabic layouts the eye usually starts on the right, so plan for that. Do not ask the generator to write Arabic text inside the image: Arabic lettering in most generators is still unreliable at the time of writing (September 2026). Generate the image without text and add the text and logo in your design software.

12. Aspect ratio and resolution

Set the aspect ratio at the start, because changing it later crops the composition: 4:5 for vertical Instagram posts, 1:1 for square, 9:16 for stories and vertical video, 16:9 for website covers and presentations. For print, state the size and plan to upscale after generation.

13. Exclusions

What must not appear: text, other logos, faces of real people, identifiable children, elements that clash with the audience's values, competing products. Written exclusions save many rounds of generation.

14. Review criteria

How will you decide the image is acceptable? Write it before generating, not after, so you do not fall for a beautiful image that does not serve the purpose. A ready-made criteria table is at the end of this guide.

The blank brief

Visual brief

Project / brand: [ ]              Brief date: [ ]          Image no.: [ ]
Purpose and use: [where the image appears; what must catch the eye first]
Audience: [who will see it]

Subject: [the object or scene, with details that must not change]
Action / pose: [what is happening]
Environment: [setting and background]
Composition: [subject position, shot size, what stays empty]
Camera: [angle] / [lens] / [depth of field]
Lighting: [source, direction, character]
Materials and texture: [required materials] / Excluded: [materials off-brand]
Brand colours: dominant [ ] / secondary [ ] / accent [ ]
Style (fixed style sentence): [ ]
References: [ref 1: what we take from it] [ref 2: what we take from it]
Text placement: [where, and how much space] - text added in design, not
generated
Aspect ratio and resolution: [4:5 / 1:1 / 9:16 / 16:9 / print size]
Exclusions: [text, logos, real faces, ...]
Review criteria: [3 to 5 criteria]
Approved by: [role]

Purpose and use: where will the image appear, and what must be seen first?

Subject and its fixed details

Composition, camera and lighting

Materials, brand colours and style sentence

Text placement, aspect ratio and exclusions

Review criteria

Turning the brief into an image prompt

Once the brief is complete, arrange its fields into a single prompt with this template. The order is not sacred, but subject first, then details, then style works well with most generators. Many generators follow English more precisely; if Arabic prompts give weaker results in your tool, translate the prompt and keep the brief in Arabic for the team.

Image prompt from the brief

[subject with details], [action or pose], in [environment].
Composition: [subject position and shot size], with calm empty space at
[text placement].
Camera: [angle], [lens], [depth of field].
Lighting: [source, direction and character].
Materials: [materials and texture].
Colours: [dominant], [secondary], with a touch of [accent].
[fixed style sentence].
No text, letters or logos in the image. [other exclusions].
Aspect ratio: [ratio].

If your tool accepts a reference image, attach the real product photo and ask it to preserve the shape and label. And if the product is real and important, consider editing rather than generating: photograph the product with your phone, remove the background and place it in a new scene, instead of asking the generator to invent its shape.

Three completed briefs

Brief one: a hero product image for a bakery

Illustrative example

Brand: Al Zaytouna Bakery, Tripoli. Purpose and use: hero image for an Instagram post announcing the whole-wheat za'atar loaf; the viewer must see the loaf and its crumb first. Audience: neighbourhood families.

Subject: a round whole-wheat loaf with a golden-brown crust and a light coating of green za'atar and sesame, a quarter cut away to reveal a soft crumb with small holes. Pose and environment: the loaf on an olive-wood board on a kitchen table, with a small dish of olive oil and a sprig of fresh thyme beside it.

Composition: the loaf in the left third, medium close-up, the cut quarter facing the camera; the upper right third relatively empty for a short headline. Camera: 45 degrees from above, 85 mm lens, shallow depth of field. Lighting: warm morning light from a window on the left, soft shadows, a hint of steam suggesting freshness.

Materials: raw olive wood with visible grain, crumpled beige linen under the board, earthenware for the dish. Excluded: glossy surfaces, plastic. Colours: dominant: warm bread brown and beige; secondary: dark olive green; accent: olive-oil gold.

Style sentence: "Realistic food photography in warm natural light, calm earthy colours, clear texture detail, an honest homely feel, no oversaturation." References: a real phone photo of the loaf (we take only its shape and crust); an earlier photo from the bakery's account (we take only the warm light).

Text placement: upper right, about a third of the width; headline and logo added in design software. Aspect ratio: 4:5. Exclusions: text, hands, people, any brand marks on the utensils.

Review criteria: the loaf matches the real one in shape, colour and amount of za'atar; the empty space fits two lines of clear Arabic type; no deformed objects or unnaturally floating oil; the image looks like the bakery's other posts.

Resulting prompt (shortened): "A round whole-wheat loaf with a golden-brown crust coated in za'atar and sesame, a quarter cut away showing a soft crumb, on an olive-wood board over beige linen, a small earthenware dish of olive oil and a sprig of thyme beside it. Loaf in the left third, calm empty space upper right. 45-degree angle, 85 mm lens, shallow depth of field. Warm morning window light from the left, soft shadows, a hint of steam. Realistic food photography in warm natural light, calm earthy colours... No text or logos. Aspect ratio 4:5."

Brief two: an awareness illustration series for a clinic

Illustrative example

Brand: Nabd dental clinic. Purpose and use: three images for awareness posts about children's dental care; this brief covers the first, "brushing before bed". Families should feel calm, not afraid. Audience: parents of children aged 4 to 10.

Subject: a cartoon child with generic, non-realistic features standing at a sink holding a small toothbrush, a parent standing beside them smiling. Environment: a simple home bathroom, a round mirror, a toothbrush cup, a wall clock showing eight o'clock.

Composition: the two characters in the left half, medium shot; the right half a calm-coloured wall for a headline and a short tip. Camera: the child's eye level, flat perspective without distortion. Lighting: soft flat lighting without strong shadows, a calm evening feel.

Materials: simple untextured surfaces, shapes with rounded edges. Colours: dominant: calm dark blue and cream; secondary: teal; accent: pale yellow for the clock and small details.

Style sentence: "Flat illustration with soft edges, simple rounded shapes, a palette of dark blue, teal and cream, simple background, no strong gradients, a calm and reassuring feel, no text in the image." Pasted word for word into images two and three.

References: the clinic's colour palette (colours only). Text placement: almost the whole right half. Aspect ratio: 1:1. Exclusions: frightening medical instruments, blood, realistic faces, any lettering.

Review criteria: the characters do not resemble real people; the number of fingers and hands is correct; the brush sits sensibly in the child's hand; colours stay within the palette; placed side by side, the three images look like one brand; the accompanying health tip is reviewed by the clinic's dentist before publishing.

Brief three: a website cover for a fitness centre

Illustrative example

Brand: Lamsa Fitness Centre. Purpose and use: the main cover image on the website home page, with a headline and a "Book a trial class" button over it. It should suggest a comfortable place, not punishing exercise. Audience: office workers new to exercise.

Subject: a small, clean studio in the morning with two yoga mats, light dumbbells arranged on a wooden shelf and a large plant in the corner; a woman in modest sportswear stretching her arms, her back half-turned from the camera and her face indistinct. Environment: a sand-coloured wall, large windows on the left.

Composition: wide shot, the figure and objects in the left third, the right two thirds calm wall and floor for a two-line headline and a button. Camera: eye level, 35 mm lens, medium depth of field. Lighting: natural daylight from the windows, soft and diffused, no harsh shadows.

Materials: light wood, matte rubber flooring, cotton fabric. Excluded: shiny chrome, mirrors reflecting the camera. Colours: dominant: sand and warm white; secondary: sage green; accent: muted orange on one mat.

Style sentence: "Realistic interior photography in natural daylight, calm warm colours, a tidy and comfortable space, a quiet morning feel, low contrast, no text or logos." References: real photos of the centre's studio (we take the layout and wall colour) so visitors do not see a place different from the one they will find.

Text placement: the right two thirds, with enough space below for the button. Aspect ratio: 16:9 for desktop, checking that a 4:5 crop for mobile keeps the figure visible. Exclusions: a clear face resembling a real person, exaggerated body proportions, text.

Review criteria: the space resembles the real studio closely enough that visitors will not feel misled; hands and fingers are correct; white or dark text is legible over the empty area; the image still works after the mobile crop. If the first criterion fails, it is better to photograph the real studio and enhance the photo than to generate.

Review criteria: the scoring table

Score each image from 1 to 3 on each criterion (1 = not met, 2 = partly met, 3 = met). Any "critical" criterion scoring 1 means the image is rejected, whatever the total.

Criterion	Check question	Critical?
Product fidelity	Does it match the real product in shape, colour, label and detail?	Yes
Integrity of details	Are hands, fingers, faces and objects free of distortion?	Yes
Rights and ethics	Is it free of real faces, other brands' logos and anything needing permission?	Yes
Purpose	Does the intended element catch the eye first?	No
Identity	Are colours and style on-brand, and does it look like previous images?	No
Text space	Does the empty area fit the required text and keep it legible?	No
Ratio and resolution	Is the ratio right, the resolution sufficient, and cropping to other sizes possible?	No

Edit instead of regenerating

Change one element at a time: if only the lighting is wrong, edit the lighting line alone, regenerate and note the effect.
Local edits: many tools let you select an area and change what is in it (remove an object, fix a hand) without touching the rest.
Extend the frame: if the image is good but there is not enough room for text, extend the canvas rather than regenerating.
Upscale: for print or large screens, upscale at the end and inspect the details after upscaling.

The names of these functions and where to find them differ between tools and change with updates, but the logic stays the same.

Common mistakes

Mistake: a two-word prompt followed by random attempts. Fix: fill in the brief first, then translate its fields into one ordered prompt.
Mistake: asking for Arabic text inside the image. Fix: generate without text and with empty space, then add text and logo in design software.
Mistake: beautiful images that do not look like each other. Fix: a fixed style sentence pasted word for word into every prompt, and fixed colour references.
Mistake: generating a real product from a description and getting a different shape. Fix: use the real product photo as a reference or start from it with edits, and reject any image that does not match.
Mistake: forgetting the aspect ratio until the end. Fix: set it in the brief, and check that crops to other sizes do not cut the subject.
Mistake: generating faces that resemble real people or imitating other brands. Fix: list it under exclusions, and stop immediately if anything like it appears.

Checklist before and after generating

I wrote the purpose, the use and the audience.
I described the subject with the details that must not change.
I set the composition, camera and lighting.
I wrote the materials and the brand colours with their roles.
I wrote the fixed style sentence and saved it for future images.
For each reference, I stated what we take from it.
I set the text placement and asked for an image without text.
I set the aspect ratio and the resolution needed.
I listed exclusions, including real faces and other brands' logos.
I scored the image with the criteria table, and no critical criterion scored 1.
The named approver signed off the image before publishing.