Develop a visual intention

How can you guide the eye in a generated image by adjusting framing, lighting and composition?

1250 € per session, excl. VAT

In short

This half-day session of 3 hours explores how framing, lighting and composition change the perception of an image. Working from a fictional case, participants formulate visual instructions, compare several outputs and compile an annotated image board with a quality assessment grid and precautions for sharing.

Why this training?

An attractive image can miss its objective if the subject blends into the background or if the lighting conceals its material texture. Identifying these flaws helps move beyond vague judgements such as « more beautiful » or « more professional ». Image generation allows different intentions to be explored, but an instruction does not provide perfect control over the output. Observing discrepancies, changing one request at a time and documenting choices avoids selecting solely on the basis of an immediate impression. Running case study: The fictional workshop « Avelune Céramique » is preparing a presentation visual for an imaginary stoneware cup. The image must evoke a calm morning, make the material texture clearly visible and leave space for a title to be added later, with no people or recognisable brands. Transfer to the workplace: For your next visual requirement, reuse the grid to describe the intention before generating images. Compare two proposals, annotate differences in framing and lighting, then have the rights and notices checked before any sharing.

Objectives of this training

  • Identify the main subject, viewpoint and areas of breathing space in an image.
  • Describe lighting in terms of its direction, hardness and apparent colour temperature.
  • Formulate an instruction combining framing, lighting and the arrangement of elements.
  • Compare variants against explicit visual criteria.
  • Identify precautions relating to rights and transparency before sharing.

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Communications staff who are new to AI-assisted visual creation.
  • Content creators wishing to acquire simple guidelines for reading images.

Prior skills

  • No prior AI knowledge required.
  • Ability to use a browser and save an image to a folder.

What you will work on

  1. Read an image before requesting one

    Objective: Connect the visual's intention to observable choices. Wide and close-up views, viewing angle, main subject, background and negative space; limitations of AI interpretation of instructions.

  2. Vary framing and composition

    Objective: Observe how the placement of the subject guides the eye. Centring, the rule of thirds as an optional guide, balance of visual masses, leading lines and space reserved for text.

  3. Explore lighting and material

    Objective: Distinguish between lighting atmosphere and how clearly the object can be seen. Diffuse or hard light, side lighting, cast shadows, warm or cool colour cast; possible inconsistencies in reflections.

  4. Evaluate and prepare for responsible sharing

    Objective: Justify the final choice and highlight points requiring attention. Score clarity, composition and lighting consistency; rights relating to reference materials, consent, applicable notices and the risks of deceptive deepfakes.

What you take away

  • Visual intention sheet for the fictional cup.
  • Instruction log and board of annotated variants.
  • Selected visual with a justification for the choice.
  • Completed quality assessment grid and pre-sharing checklist.

Method and assessment

  • Short instructional inputs followed by immediate practice using the running case study.
  • Individual work followed by work in pairs, with cross-review of outputs.
  • Guided group discussion: what works, what needs correcting and what remains a human decision.
  • Preparation: Arrange access to a tool authorised by your organisation; check the available generation and saving options. Prepare a folder of fictional or anonymised images that are authorised for use, with no confidential data or identifiable portraits.

Observable criteria: The main subject remains identifiable and the space reserved for the title is clear. ; The annotations distinguish between light direction, shadows and the visibility of the material texture. ; Each comparison states the requested parameter and the changes actually observed. ; The grid assigns each visual criterion a score from 0 to 2 with justification. ; The sharing checklist documents rights, any consent required and notices to be reviewed..

Frequently asked questions

Do you need to be able to draw?

No. The exercises are based on observing and describing simple visual choices.

Is a specific software package required?

No, only a tool authorised by your organisation, depending on the access available.

Can you use a real portrait?

The exercises exclude this. Outside the session, check consent, rights and the risk of impersonation associated with deepfakes.

Can the visual be published immediately?

It remains subject to internal approvals, terms of use and generation disclosures applicable to the context.

Last updated: