AI on site, without engaging your own liability
Minutes, progress statements and claims too often get written at night, after the day on site. AI can take over that load, as long as contract documents never end up in a cloud with no guarantees.
Site managers, works supervisors, project leads in general contracting: your day happens on site, and the paper trail catches up at night. Yet that trail, signed minutes, the site diary, the defect list, is what counts the day a delay or a defect is disputed. Pasting a drawing or a progress statement into a consumer AI hands your client’s contract documents to a tool that has promised nothing.
The point is not to push AI aside: your teams and subcontractors already use it. It is to frame it, for documents that are complete, dated and sent on time, without ever feeding it what must stay inside the file.
- Evenings lost to minutes, reports and progress statements after the day on site
- Delays and defects poorly documented, impossible to defend when the final account is disputed
- Drawings and contract documents exposed in consumer AIs with no data processing agreement
- Delay and extra-cost claims weakened because nothing was put in writing in time
See for yourself
A prompt structured in 5 blocks: Instruction · Role and audience · Data · Constraints · Format. An unframed AI invents; the frame stops it.
In a workshop, these prompts are adapted to your sites: your minute templates, your SIA requirements, your clients. One rule never changes: no contract document ever touches a public AI.
Procès-verbal de réunion de chantier hebdomadaire
The weekly minutes sent out the same evening, with decisions, owners and deadlines in black and white.
See it in Chaptal →Recadrage d’un sous-traitant en retard ou en défaut
A firm, factual letter that states the facts and the deadlines without burning the relationship.
See it in Chaptal →Réclamation de prolongation de délai et de surcoût
A structured claim, backed by a dated chronology, that stands up in front of the client.
See it in Chaptal →30 use cases cover your line of work in the Chaptal catalog, out of 750 in the corpus · each one states what it produces, what it expects from you, and what it refuses to do.
« Le Collaborateur, augmenté »
Your site managers learn to draft with AI on their own site documents, without exposing anything.
See the training →AI audit & advisory
A review of your current AI usage, site by site, and a written framework for your teams.
See audit & advisory →Foire aux questions
Can we paste a drawing or a contract document into a public AI?
Not without written guarantees. A consumer AI with no data processing agreement may keep and reuse what it is given. We work with anonymised data or with tools that provide those guarantees.
Do minutes prepared with AI carry the same weight as handwritten ones?
Yes. What counts is the signature and the distribution, not the drafting tool. The AI prepares the text from your notes; you review, correct and remain the author of the document.
Our site managers are not IT people, is that a problem?
No. A well-built prompt works like a form: you provide your site notes, it returns the document in your format. Training runs on your own cases, not on abstract examples.
Does AI know the SIA standards and Swiss construction law?
It knows the broad lines and can get the details wrong. Our prompts require it to flag what must be checked in the standard itself instead of asserting it.
Where do we start on a project already under way?
With the documents that come back every week: the site meeting minutes, the next-day coordination message, the defect follow-up. Those free up hours right away, without changing how you work.