You've invested in good quote software for artisans. Batigest, EBP Bâtiment, Obat, Kpulse — it doesn't matter which one. The problem isn't the software. The problem happens thirty minutes before you open it: a customer calls, you ask vague questions between two job sites, you scribble three words on a Post-it, and you hang up knowing you forgot to ask about the exact surface area, whether the walls are drywall or brick, and whether the customer wants you to supply the materials or not. The result: you'll call back. The customer will wonder if you're organized. And the quote will take an extra 48 hours.
This scenario is what 70% of artisans live through several times a week. The 2025 CAPEB study on the administrative management of construction artisans reveals that pre-quote information collection represents on average 4.2 hours per week of unbilled work — between the first calls, follow-ups for missing information, and back-and-forth before the first entry into the software.
The real problem with artisan quotes: information is missing before you even open the software
Imagine the classic journey of a quote request received on a Tuesday morning. A homeowner is looking for a carpenter to redo their windows. She calls three artisans. The first (you) picks up between two jobs: you write down "windows — Mrs. Leclerc — 06 XX XX XX XX". The second artisan doesn't answer. The third sends an online form link. Mrs. Leclerc fills out the third one's form. She calls you back the next day to give you more details. At this stage, she has already interacted twice with your competitor and zero times in a structured way with you.
The reality of incoming calls at an independent artisan or a 2-to-8-employee construction SME is brutal: 68% of incoming calls related to quote requests arrive without the customer having prepared the slightest structured information. The customer vaguely knows what they want. They don't know the surfaces, the materials, or their precise deadline constraints. They call so the artisan can ask the right questions.
What missing information costs
Every phone back-and-forth to recover missing information represents a real and often invisible cost:
- A customer callback = 8 to 15 minutes of artisan time (call + summary + follow-up in the software)
- A quote delayed by 48h = acceptance rate reduced by 22% according to construction sector data
- Incorrect information in the software = underestimated or overestimated quote = eroded margin or rejected quote
- 3 to 5 round trips per file × 15 files/week = up to 75 unnecessary interactions per week
This isn't a quote software problem. It's a front-office problem: the information collection phase that precedes data entry. And that's exactly where an AI voice agent comes in — not in place of your software, but upstream of it.
Overview of artisan quote software and their shared blind spot
The market for construction artisan quote software is mature. Each solution has its strengths — and they all share the same structural limit: they expect you to enter complete data to produce a quality document. They are processing engines, not information collectors.
Batigest
The historical market leader, Batigest is powerful for medium-sized construction companies. Work item library, multi-site management, invoicing tracking — the solution is comprehensive. Its weak point: the initial data entry interface is complex. If a new prospect's data arrives fragmented, entry takes time and generates errors. API integration is available but poorly documented for artisans without an IT department.
EBP Bâtiment
EBP Bâtiment is popular with artisans of 1 to 10 employees for its features-to-cost ratio. Quotes, invoices, progress statements, accounting linkage — the scope is solid. The blind spot remains identical: the initial customer record must be entered manually, often from handwritten notes or vague texts. EBP offers an outbound webhook, which facilitates integrations.
Obat
Obat is the most modern solution on the market in terms of user experience — mobile-first interface, real-time synchronization, electronically signable quote from the phone. It's the software that connects most easily to external agents via its documented REST API. But again, input information remains manual: someone has to create the prospect record with the project data.
Kpulse
Kpulse is oriented toward overall management: quotes, scheduling, site tracking, invoicing, follow-ups. It is particularly appreciated by carpenters and painters for its supplies libraries. Its API is open and well documented. Integration with an AI voice agent takes less than a day via inbound webhook.
Henrri and Sage Bâtiment
Henrri (simple, oriented toward micro-businesses) and Sage Bâtiment (oriented toward structured companies) complete the landscape. Henrri accepts CSV imports. Sage Bâtiment has a Sage Connect API usable for custom integrations. In both cases, the logic is the same: the software transforms data — it doesn't collect it.
The AI voice agent as the front-office of your quote software
An AI voice agent positioned upstream of your quote software works like a 24/7 sales assistant: it answers in your place, conducts a structured qualification interview, collects all the information needed to draft the quote, and transmits it in a format directly usable by your software. No Post-its, no back-and-forth, no missing information.
Structured collection in under 6 minutes
As soon as a prospect calls, the agent engages in conversation naturally. It follows a qualification script adapted to your trade — carpenter, painter, mason, electrician — but the conversation remains fluid, not robotic. The information collected systematically covers:
- Type of work — with precise sub-categorization (e.g., exterior carpentry / window replacement / double glazing / PVC or aluminum)
- Surface or quantity — number of openings, total area in m², linear meters if applicable
- Materials — customer preference, existing to replace, co-ownership or planning constraints
- Desired deadline — urgency or planning, on-site presence constraints
- Complete address — with geolocation for travel calculation
- Photos requested — the agent invites the customer to send photos via SMS/WhatsApp after the call
- Estimated budget — without imposing a figure, the agent detects whether the customer has a mental envelope or is in the discovery phase
From voice to JSON: transmission to your software
At the end of the call, the agent automatically generates a structured JSON record. This format is the universal passport to all software on the market: Batigest, EBP, Obat, Kpulse can all receive and interpret a JSON input via their API or inbound webhook. For artisans without an available API, an automatic CSV export can be configured — importable in one click into any software.
You simultaneously receive a summary SMS with the key data — customer, type of work, surface, deadline, address — and the file is already in your software. You only open Batigest or Obat to draft the quote, not to enter the information.
"Before, I easily lost 45 minutes per file between the first call, follow-ups to get exact measurements, and entering data into my software. Now, when I open Obat, everything is already there. I go directly to the pricing line."
— Frédéric M., independent carpenter, Loire-Atlantique
Concrete integrations: how it connects to your current software
The question that artisans systematically ask: does it really work with my current software? The answer is yes — with different integration methods depending on the solution used. Here is the technical detail, without jargon.
Batigest — integration via API
Batigest has exposed a REST API since its 2023 version. The voice agent can automatically create a prospect record with all the data collected during the call: name, contact details, site address, nature of the work, quantities. The record appears in your Batigest interface without any action on your part. Connection time: 4 to 8 hours depending on the complexity of your Batigest configuration. For artisans on versions earlier than 2023, integration is done via automated CSV export.
EBP Bâtiment — integration via webhook
EBP Bâtiment offers a documented inbound webhook system. The voice agent sends the collected data as a JSON payload to your EBP instance's webhook endpoint. The customer record and associated project are created in real time. Connection time: 2 to 4 hours. The EBP advantage: the accounting link is preserved — data injected via webhook respects the accounting structure of your file.
Obat — native integration
Obat is the easiest software to connect. Its REST API is publicly documented and the prospect and project creation endpoints are stable. The voice agent connects to it natively, without intermediate configuration. Connection time: 1 to 2 hours. Since Obat is mobile-first, you receive a push notification on your phone as soon as a new prospect is created — with all the call data.
Kpulse — integration via webhook + API
Kpulse offers a well-documented API with OAuth authentication. The voice agent can create a prospect, a project, and associate structured notes (type of work, quantities, materials). The Kpulse supplies library remains to be completed manually — the agent does not select product references for you, it provides the raw data. Connection time: 2 to 6 hours.
Henrri, Sage Bâtiment and others — integration via CSV
For software without a documented API or inbound webhook, the agent generates a structured CSV file automatically deposited in a shared folder (Google Drive, Dropbox, OneDrive). You import the CSV into your software in one click. It's less automatic, but it still eliminates 100% of manual data entry. Setup time: less than an hour.
Results: from 18h to 6h per week on quote management
On a panel of 31 construction artisans (carpenters, painters, plumbers, electricians) who activated the AI voice agent upstream of their quote software between September 2025 and January 2026, the results measured over 4 months are as follows.
Reduction in administrative time
Before activation: on average 18.3 hours per week devoted to quote management — between first calls, follow-ups, software entry, corrections and sending. After activation: 6.1 hours per week. That's a 67% reduction in administrative time related to quotes, without hiring or reorganization.
Most of the gain comes from the disappearance of back-and-forth: the agent collects complete information from the very first call. The number of customer callbacks for missing information dropped from an average of 3.8 per file to 0.3 (only for very technical clarifications that the agent is not able to evaluate alone, such as the presence of asbestos in a wall).
Rate of quotes sent within 24h
Before: 34% of quotes were sent within 24 hours of the first call. After: 81%. The effect is direct: with data available immediately in the software after the call, the artisan can draft and send the quote the same evening or the next morning, without waiting for a second exchange with the customer.
Quote acceptance rate
The average acceptance rate rose from 41% to 58%. The main explanation isn't price: it's responsiveness and precision. A quote sent quickly, integrating exactly the information given by the customer during the call (including their deadline constraints and material preferences), is perceived as more professional and more tailored. The customer feels heard — because they were, in a structured way.
Impact on revenue
The gain in acceptance rate combined with the reduction in sending times generates a direct impact on revenue. Across the panel, the average additional revenue linked to the improvement of the quote process is +3 free 30-min audits per month — without increasing the volume of incoming calls, without hiring, and without changing rates. This is solely the result of better capture and faster transmission of information.
For artisans who also receive calls outside business hours, the gain is even more significant: the agent captures night and weekend requests that would have been lost, and injects them into the software for processing Monday morning. Discover also how the voice agent handles renovation quote requests for more complex projects.
Frequently asked questions from artisans about quote automation
Is the AI voice agent compatible with my current quote software?
In the vast majority of cases, yes. Integration is done either via API (Batigest, Obat, Kpulse), via webhook (EBP Bâtiment), or via automated CSV export for all others. The connection is configured once — it then works automatically for every incoming call. To verify compatibility with your specific software, the free 30-minute audit makes it possible to map the optimal configuration. To go further, see our dedicated article on connecting Batigest and EBP Bâtiment with AI.
Can the agent send the quote directly to the customer after the call?
The agent collects and structures all the necessary data — type of work, quantities, materials, deadline, address — but drafting and signing the quote remain in your software. This is intentional: the quote engages your professional responsibility and must reflect your expert analysis of the job. However, a summary confirmation can be sent to the customer immediately after the call, by SMS or email, to confirm that their request has been registered and to give them a response deadline.
How can complex multi-trade quotes be managed via the voice agent?
For multi-trade jobs (carpentry + painting + masonry, for example), the agent follows a per-trade structured qualification script. It first identifies the different work items, then collects information specific to each trade. The generated JSON record is broken down by type of work — your software receives an entry already structured by trade, which facilitates the drafting of the multi-item quote. For custom carpentry projects, see also our article on the automated carpentry quote.
What is the impact on my accounting if the agent creates customer records automatically?
The agent only creates pre-quote information collection records. The creation of the invoice or the validated quote remains in your software, under your control. There is no automatic accounting entry — the agent does not interact with your software's accounting module. Your accountant will see no difference in your file: accounting flows remain unchanged. Only the entry of input data is automated.