CCTP specification by trade with ChatGPT: method and checks
CCTP by trade with ChatGPT: distinguishing retained, removed and new work in renovation, citing the applicable DTU, and the method.
By Educasium
Install the ChatGPT skill: CCTP specification by trade →
Only the painting trade is requested for the renovation of a building's lobby. The drawings clearly identify the area concerned, but neither the existing substrate nor the expected finish are specified at this stage. An architect asking ChatGPT to draft that trade's CCTP has two ways to handle the gap: prescribe a plausible brand and shade so the document looks finished, or draft the trade actually requested while leaving those two details in a list to confirm. The second option takes more effort to defend to the client; it is nonetheless the only one that protects the project from an invented requirement no document actually supports.
This article details what a CCTP must inventory before a single sentence is drafted, why a renovation requires distinguishing retained, removed and new work, the practice that avoids ambiguity over execution rules — explicitly citing the DTU applicable to each service —, and the method for drafting this document with ChatGPT without ever importing a requirement foreign to the project.
Summary
- What a CCTP must inventory before any sentence is drafted
- In renovation, three categories that must never blur together
- Why explicitly citing the applicable DTU for every service matters
- Drafting a CCTP by trade with ChatGPT, step by step
- When this pack is not enough
- Training to make CCTP documents more reliable
- Frequently asked questions
What a CCTP must inventory before any sentence is drafted
The particular technical specifications (CCTP) prescribe, trade by trade, the work items, their installation and the expected checks. Its reliability depends directly on the inventory carried out beforehand: affected items, precise locations and the source of every figure carried into the text. A CCTP drafted without that prior inventory often reproduces, unintentionally, the requirements of a generic template rather than those of the real project.
A template remains a template until checked against the project
The difference between a reliable CCTP and a generic one is not visible in the writing style, but in the correspondence between every article and a project document that justifies it. A CCTP template explicitly identified as such, then checked item by item against the real drawings and brief, produces a reliable document; the same template copied without that check silently carries over requirements unrelated to the site at hand.
In renovation, three categories that must never blur together
A renovation CCTP that does not explicitly distinguish what is retained, what is removed and what is new exposes the contractor to an ambiguous reading of their actual scope of work.
| Category | What it designates | Risk if not explicitly distinguished |
|---|---|---|
| Retained | Existing item kept as-is, outside the scope of works | The contractor mistakenly prices a service that was not requested |
| Removed | Existing item taken out before new work begins | The CCTP omits the removal service, though it is required to install the new work |
| New | Item created or replaced under the contract | A new-work requirement mistakenly applied to a retained item |
Why this distinction also protects the budget
A removal service forgotten because the CCTP presents the new work as if it were built from scratch almost always turns into a change order mid-site, priced through negotiation from a position of weakness rather than through competitive bidding at the tender stage. Distinguishing the three categories from the outset avoids this situation, far more common than a dimension or material error.
Why explicitly citing the applicable DTU for every service matters
The unified technical documents (DTU) describe the expected execution rules for each trade. As with an administrative standard such as NF P03-001, contractual certainty comes from explicit citation: a CCTP that prescribes an installation method without ever naming the reference DTU leaves ambiguity over the actual level of quality expected, ambiguity that rarely favours the client if a defect appears later.
What this actually changes for the consulted contractor
A contractor responding to a CCTP that explicitly cites the applicable DTU for every service prices on a clear basis, with no need to guess the expected level of quality. A CCTP silent on this point, conversely, exposes the same article to two different readings by two consulted contractors, which later complicates comparing the bids received for that trade.
Drafting a CCTP by trade with ChatGPT, step by step
Step 1: Inventory items, locations and sources before drafting anything. This inventory, done before asking ChatGPT anything, guarantees every requirement drafted afterwards can be linked to a real project document.
Step 2: For renovation, explicitly distinguish retained, removed and new work for every item. This distinction must appear in the document's very structure, never only in an isolated sentence easy to miss on a read-through.
Step 3: Structure scope, interfaces, services, installation and checks based on data actually available. A section with no available data must stay flagged as incomplete, never filled in with a plausible requirement just to look exhaustive.
Step 4: Explicitly cite the DTU or technical reference applicable to every service. This citation must be verified before use, never carried over automatically from a template without confirming its relevance to the item concerned.
Step 5: Link every article to the matching DPGF items, without importing requirements foreign to the project. A CCTP article matching no DPGF item signals either a pricing omission or a requirement carried over from another site that should be removed.
Fictional example for testing the method: only painting is requested; drawings identify the lobby but not the substrate or finish. The correct request to ChatGPT is to draft the requested trade while leaving those details unconfirmed, never prescribing an arbitrary brand or shade just so the document looks finished.
When this pack is not enough
A well-structured CCTP does not resolve every situation encountered when drafting a technical trade.
A complex interface between two trades — an item at the boundary between structural work and waterproofing, for example — often requires a design decision between the two trades before drafting, rather than an automatic split decided by the CCTP text alone. A technical requirement whose regulatory compliance remains uncertain — a new construction system, evolving regulations — must be checked separately with an engineering consultant or a technical inspector before being added to the final document. A CCTP prepared with AI assistance, ChatGPT like any other tool, remains in every case a draft: it must be reviewed and approved by the architect or engineering consultant responsible for the trade before it is circulated to consulted contractors.
Training to make CCTP documents more reliable
A well-designed pack is not enough on its own: it also needs adapting to every trade, with items and interfaces that change from one project to the next. For an architect practising as a self-employed professional (NAF code 7111Z in France), the FIFPL fund covers part of the cost of Qualiopi-certified training under 2026 criteria set at €300 per day and €900 per year, with e-learning capped at 50% of the daily rate — funding that must be arranged before training starts, not after.
Frequently asked questions
Should a DTU always be cited in a CCTP, even for a simple service?
It is recommended as soon as the service has a reference DTU, to set a clear level of quality shared by every contractor consulted. Not citing one does not invalidate the CCTP, but leaves room for interpretation that two contractors may fill differently. That reference must be checked for every service before the document is circulated, rather than copied out of habit from a template used on a different site.
What should be done if the drawings do not specify the existing substrate of an item to be renovated?
That detail should stay flagged as unconfirmed in the CCTP, never filled in with a plausible assumption just so the document looks complete. A site visit or a sample opening remains the only reliable way to resolve that uncertainty before final drafting. The price attached to that service must then remain conditional in the corresponding DPGF, until the actual substrate is confirmed.
Can ChatGPT decide on its own how to split an item between two trades?
No: ChatGPT can flag an ambiguous interface between two trades, but the actual split is a design decision between the trades concerned, to be validated by the architect or engineering consultant before it is added to the CCTP. Until that decision has been made, the item concerned must stay flagged in both CCTPs as an interface to clarify, rather than defaulted to one of the two trades. Once the split is confirmed, it must be carried through identically into the CCTP and DPGF of both trades concerned, to avoid a service ending up neither priced nor executed.
Why does distinguishing retained, removed and new work change the trade's final price?
Because a removal service forgotten for lack of that explicit distinction almost always turns into a change order negotiated mid-site rather than a price set through competitive bidding at the initial tender — a price difference that has nothing to do with the quality of the contractor chosen. That distinction must therefore appear explicitly in every relevant CCTP article, rather than being implicitly inferred from the general description of the item. Once established, it must be carried through identically into the matching DPGF, so every removal service gets its own price line from the tender stage onward.
A reliable CCTP rests on three elements that do not change from one trade to the next: an inventory verified before any drafting, an explicit distinction between retained, removed and new work in renovation, and a technical reference cited and verified for every prescribed service.
Once the CCTP is drafted, our ChatGPT method for DPGF breakdown by trade details how to price the services it prescribes, and our ChatGPT method for DCE consistency checking revisits the cross-check against the tender file's other documents.
Training 100% fundable via OPCO/FIFPL. Qualiopi-certified programme. To structure your technical requirements with AI as part of our AI for architectural management training, contact Educasium and specify your status (employee, self-employed, business owner) and your goal. This pack is also available on its dedicated ChatGPT skill page.