Est-ce que RestorFX dispose d'une garantie ?

Oui, RestorFX est un traitement permanent. Nous offrons une garantie de tenue dans le temps de 10 ans lorsque le traitement est appliqué sur une peinture d'origine (n'ayant pas été repeinte ou reçu de traitements préalables). Cette garantie couvre le fait que le traitement ne s'écaille pas et ne délamine pas.

Combien de temps dure le traitement RestorFX ?

RestorFX est un traitement permanent. Le véhicule est généralement immobilisé 24 à 48 heures pour le traitement complet.

Quelle est la différence entre polissage/lustrage et RestorFX ?

Le polissage retire une couche de vernis (abrasion). RestorFX est un procédé technique et chimique en 13 étapes qui fusionne avec le vernis pour le réparer durablement, sans devoir repolir régulièrement.

Est-ce moins bien qu'une nouvelle peinture ?

Ni mieux ni moins bien : cela dépend du dommage. RestorFX conserve la peinture d'origine, évite les écarts de teinte, et c'est plus rapide et souvent moins coûteux. La peinture s'impose en cas de dégâts majeurs.

Différence entre RestorFX et traitement céramique ?

RestorFX corrige les défauts (rayures, oxydation). La céramique protège et facilite l'entretien. Les deux sont compatibles et complémentaires.

Entretien après le traitement RestorFX ?

Évitez toute action chimique ou mécanique pendant 2–3 semaines (pas de rouleaux, pas d'essuyage). Un simple rinçage au jet est recommandé durant cette période.

Copilot Illuminator

Phase 1 · Define the situation

Situation Appraisal

Narrative: Imagine this: you have 240 SQV reports, each one in PDF format, each one written slightly differently, and you need to decide quickly which sites are ready. The issue is not that we lack data. The issue is that the data is trapped inside unstructured documents.

Manual review is slow, difficult to compare, and vulnerable to missed risks. The goal of this scene is to move from document chaos to a structured decision table that can support a clear site-selection discussion.

Learning focus

Clarify the decision question.
Identify the unstructured source: SQV PDFs.
Prepare the Excel output structure before using AI.
Frame the business question: which sites are ready, which need follow-up, and which should not be selected?

Phase 2 · Use AI as an alternative

Analysis & Alternatives

Narrative: Now we move from problem definition to analysis. This is where we introduce Copilot in Excel. The learner opens Copilot, attaches the SQV PDF report package, and writes a structured prompt that tells Copilot exactly what to extract, how to classify it, and how to format the result.

The key is not simply to ask, “Analyze these documents.” The key is to give Copilot a clear role, context, task, output schema, rules, and validation expectations.

Best prompt structure

Role: “Act as a clinical trial site qualification analyst.”
Context: “Use only the attached SQV PDF reports as the source.”
Task: “Extract one row per site and populate the Excel table.”
Schema: Define the exact columns: Site ID, Country, Enrollment, Outcome, Readiness Score, Risks, Gaps, Summary, Action.
Rules: Do not assume missing information. Flag unclear evidence as follow-up.
Output: Return a structured table aligned with the Excel headers.

See full video demo

Placeholder: this is where the practical end-to-end workflow demo will be linked or embedded.
The demo will show how the reports are shared, how the prompt is used, and how the output is checked in practice.

Phase 3 · Turn analysis into action

Decision & Implementation

Narrative: Now we have structured data, but this is not the moment to blindly trust the AI. This is the moment to introduce the most important control principle: human in the loop.

Copilot accelerates the review, but the study team remains responsible for validating the evidence, challenging the output, and making the final decision.

Human-in-the-loop checks

Traceability: Can each output be linked back to evidence in the SQV report?
No assumptions: Did Copilot flag missing information instead of inventing it?
Consistency: Are sites scored using the same criteria?
Risk review: Are operational, recruitment, equipment, and documentation risks correctly interpreted?
Outlier check: Does any site look unusually strong or weak and require manual verification?
Final decision: Humans approve, reject, or request follow-up based on validated evidence.

Decision rule shown

Filter sites with outcome Meets.
Prioritize high enrollment and low operational risk.
Flag sites needing documentation follow-up.

Decision Making Framework: SQV Review

Knowledge Check