Editing an engaging narrative

How can you tighten a video edit without compromising clarity, emotion or faithfulness to the message?

1960 € per session, excl. VAT

In short

A 7-hour day, excluding breaks, to diagnose an edit, rebuild its narrative progression and adjust its pacing with support from generative AI. Working from a fictional case, you produce a re-edited version, document your editorial decisions and compare its quality with that of the original edit.

Why this training?

A film can be technically polished yet lose its viewer: the stakes are introduced too late, information is repeated, transitions are confusing or the pace is uniform. Cutting more is not enough; every removal must preserve causal links and useful breathing space. Generative AI can suggest alternative structures based on a shot log and a transcript. These suggestions remain editorial hypotheses: the editor checks their feasibility, the meaning of what is said and their effect when viewed. Case used throughout: The fictional association « Avelune Réparation » is preparing a video presenting a bicycle repair workshop. Its first three-minute edit reveals the beneficiary’s problem late and repeats several explanations. A two-minute version must be produced that remains faithful to the testimonies, with an explicit invitation at the end. Transfer to work: On your next edit, reuse the diagnostic grid before making any cuts. Ask a colleague who is not involved to compare two versions, then record the corrections adopted and any points of understanding that still need checking.

Objectives of this training

• Identify breaks in understanding and overlong passages using time references.
• Build two narrative progressions from the same rushes.
• Formulate a request for editorial analysis by AI without having it invent shots.
• Adjust cuts, silences and audio transitions in a re-edited version.
• Compare versions using a grid and document the decisions made.

Who is it for and what are the prerequisites?

Roles and tasks concerned

• Editors and videographers producing institutional films, testimonials or short-form content.
• Audiovisual communications officers overseeing revisions to edits.

Prior skills

• Be able to edit, manage multiple audio tracks and export a video.
• Have previously used an AI assistant and know how to specify context, constraints and the expected format.

What you will work on

1. Diagnosing the first version

   Objective: Distinguish useful slowness from unnecessary length. Opening promise, stakes, progression, repetition and information load.

2. Preparing a reliable AI-assisted analysis

   Objective: Give the AI verifiable editorial material. Transcript, shot inventory, duration constraints; rights, consent, confidentiality, synthetic content disclosures and deepfake risks.

3. Comparing two structures

   Objective: Choose a progression based on the available rushes. Hook, reveal, demonstration, resolution; distinction between suspense and artificially withholding information.

4. Re-editing the decisive moments

   Objective: Tighten the narrative without distorting what is said. In and out points, ellipses, cutaways, continuity of movement and audio transitions.

5. Shaping audiovisual pacing

   Objective: Alternate purposeful accelerations and breathing spaces. Shot duration, silences, verbal density, music and text readability.

6. Testing and safeguarding delivery

   Objective: Assess progress without confusing preference with understanding. Cross-review viewing, identical questions, alternating presentation order and final checks.

What you take away

• Timestamped diagnosis of the original edit.
• Revised narrative plan and editing rundown.
• Two-minute export and organised editing project.
• Comparison grid, decision log and rights check.

Method and assessment

• Short teaching inputs followed by immediate practice on the case used throughout.
• Individual work followed by work in pairs, with peer review of the outputs.
• Guided group discussion: what works, what needs correcting and what remains a human decision.
• Preparation: Set up a working editing workstation and access to a tool authorised by your organisation. Consult the fictional case file provided: brief, rushes, timestamped transcript and usage permissions.

Observable criteria: Each identified weakness refers to a timestamped passage and an explicit narrative effect. ; The final version runs for two minutes, with a tolerance of five seconds. ; No cut changes the meaning of a testimony or introduces misleading causality. ; The test captures understanding of the problem, the solution and the expected action. ; The necessary rights, consents and disclosures are recorded before distribution..

Frequently asked questions

Does the AI need to analyse the video directly?

No. The transcript and timestamped log are sufficient; human viewing remains essential.

Can we work with our own rushes?

Only if they are anonymised and authorised; otherwise, the fictional case supports the entire programme.

Do we need to generate images or voices?

No. No deepfake is needed; any change of identity requires a specific review of rights and consent.

Does the test prove improved viewer retention?

No. It documents understanding during the workshop; viewer retention requires observations during actual distribution.