Your first images with guidance

How do you describe the image you want and check that the result matches your intention?

1250 € per session, excl. VAT

In short

In a half-day session lasting 3 hours discover how to turn an intention into a clear visual instruction. You write an initial prompt, compare images and adjust a description. You leave with a reusable file combining a prompt, a selected image and a checklist, with no prior knowledge of AI required.

Why this training?

A request such as ‘a beautiful gardening poster’ leaves the objects, framing and text placement unspecified. Describing these elements helps express an intention that can be checked rather than a vague aesthetic preference. An appealing image may nevertheless contain distorted objects, unwanted lettering or an unusable composition. Learning to spot these discrepancies helps you choose a targeted correction and distinguish a visual draft from an image ready for distribution. Running case study: The fictitious association ‘Avelune Botanique’ is preparing a poster for a repotting workshop. It wants a vertical illustration showing three terracotta pots, a small trowel and a bag of potting compost, with no people or brands, and a clear area at the top where the title can be added later. Application at work: For your next piece of material, reuse the template with a non-sensitive subject. Keep the initial prompt, a revision and their evaluations; have the rights and necessary notices checked and approved before any distribution.

Objectives of this training

  • Identify the subject, context, style and composition of an image
  • Write a short prompt with explicit visual constraints
  • Compare two results using a shared assessment grid
  • Revise an instruction by targeting an observed discrepancy
  • Identify the checks on rights, consent and notices required before distribution

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Communications assistants discovering how to create illustrations with AI
  • Staff who need to prepare simple internal materials
  • Association managers wishing to express their visual intentions more clearly

Prior skills

  • No prior knowledge of AI or drawing skills required.
  • You must be able to enter text and save a file on a computer.

What you will work on

  1. From idea to brief

    Objective: Describe an observable visual intention. The principle of image generation, variability of results and the distinction between illustration and photographic evidence. Review of the poster requirements.

  2. Assemble an initial instruction

    Objective: Build a simple, testable prompt. Subject, setting, descriptive style, framing, lighting and constraints. Prioritising concrete instructions; limitations in accurately reproducing quantities and embedded text.

  3. Observe and correct a discrepancy

    Objective: Evaluate the result without making multiple changes. Grid with three ratings: meets requirements, partially meets requirements, does not meet requirements. Checking the objects, their consistency, framing and the clear area.

  4. Prepare for responsible use

    Objective: Decide which checks are needed before publication. Rights to reference material and terms of use for outputs; consent from the people depicted; notices required according to the context. Deepfakes: risks of impersonation and misleading presentation.

What you take away

  • Visual brief for the Avelune Botanique poster
  • Completed prompt template and corrected version
  • Selected image with its limitations noted
  • Comparison grid and pre-distribution checklist

Method and assessment

  • Short explanations followed by immediate practice on the running case study.
  • Individual work followed by work in pairs, with participants reviewing each other’s outputs.
  • Guided group discussion: what works, what needs correcting and what remains a human decision.
  • Preparation: Arrange access to an image generation tool authorised by your organisation; check the access conditions before the session. Prepare only fictitious or anonymised descriptions, without personal photographs or confidential documents.

Observable criteria: The prompt specifies the objects, style, orientation and area reserved for the title. ; The comparison distinguishes between elements that meet requirements, are missing or are incorrect. ; The revision targets an identified discrepancy without changing the entire request. ; The final choice is based on the grid rather than personal taste alone. ; The final checklist examines rights, consent, the risk of deception and applicable notices..

Frequently asked questions

Do you need to be able to draw?

No. The exercises are based on describing objects, colours and positions.

Can you use a photograph of a colleague?

Not during the exercises. For later use, check consent, rights and internal rules before importing anything.

Does the same prompt always produce the same image?

No. Keep the results and compare how well they meet the requirements rather than expecting an identical reproduction.

Is the resulting image ready for immediate publication?

Not automatically: check for defects, review the terms of use and rights, and check the notices applicable to the distribution context.

Last updated: