The work behind the building site
For many trades and construction businesses, a lack of orders is not the problem. The problem is in the office: enquiries arrive by email, phone, web form and messenger, delivery notes pile up, incoming invoices need checking, and in the evening someone types up the day's reports. This work is necessary, but it does not generate revenue, and it often ends up on the owner's desk.
This is exactly where AI automation comes in. It does not replace the work on site, nor the experience a master tradesperson brings to pricing a quote. It takes over the recurring steps in between: reading, sorting, transferring and following up. This article shows which processes are suitable and what you should look out for.
Receiving and pre-sorting enquiries
An enquiry in the trades is rarely complete. Measurements, photos, the address or a preferred date are missing. Before anyone can even price the job, several messages often go back and forth. An AI agent can read incoming enquiries, whether they arrive via the web form, by email or as a message, and put them into a consistent format.
What the agent recognises
- Which service is requested, for example painting, roof repairs or electrical installation
- Location and catchment area: is the job within your area?
- Scope and urgency, as far as the enquiry reveals them
- Which details are still missing for a quote
Asking follow-up questions and requesting photos
If something is missing, the agent sends a friendly follow-up question on your behalf, for example asking for two or three photos of the surface or the damage. When the photos come back, they are attached to the right job. You decide which follow-up questions go out automatically and which you want to see first.
Enquiries that are not a good fit, such as jobs outside your area or for services you do not offer, are flagged. You decide whether a polite decline follows or whether you look at the case yourself.
Preparing draft quotes
Once an enquiry is complete, the agent can put together a draft quote from your price lists and standard text blocks. This does not replace a person doing the pricing, but it removes the writing: line items, standard wording, payment terms and the cover letter are already prepared.
Approval is what matters. A quote only leaves the business once someone has checked it and adjusted it where necessary. In the trades in particular, a price often depends on things that only become clear on site. The draft saves time; the responsibility stays with you.
Delivery notes and incoming invoices
Materials are delivered to site, the delivery note ends up in the van or the site cabin, and the invoice arrives by email weeks later. Anyone who wants to check that what was invoiced matches what was delivered has to hunt down both. This work is well suited to automation.
- Delivery notes are photographed or scanned and read, even when they carry handwritten notes.
- Incoming invoices are captured: supplier, line items, quantities, amounts.
- Invoice, delivery note and purchase order are matched, and discrepancies are flagged.
- An allocation to the building site or cost centre is suggested.
- The document and a booking suggestion are passed to DATEV (the accounting software used by most German tax advisers) or to your accounts team.
Invoices above a threshold you set go to a person for approval. You can read more about this in the article on automating incoming invoices.
Appointments, site diary and reports
Scheduling appointments
Arranging site visits takes time, especially when several messages are needed. An agent can suggest free slots from your calendar, interpret the customer's reply and enter the appointment. Short-notice changes, for example because of weather or illness, remain with you, but the agent can prepare the notifications to everyone affected.
Reports from voice notes
Nobody on a building site likes writing long texts. A voice note at the end of the day is quicker. The AI can turn it into a structured site diary entry or a report for the customer: what was done, what material was used, what is still outstanding. The report is read before it is sent, so that nothing incorrect reaches the customer.
Drawing the line: AI automation and trade software
AI automation is not trade software. It works with the programs you already use and takes over the steps between them. If, on the other hand, you lack software that covers measurement, quoting, job scheduling, photo documentation and invoicing in a single workflow, that is a different question.
For painters, roofers and electrical contractors, the Scalient Group has its own product for this: Gewerkheld. Our guide to choosing software for the trades describes what to look for in general. The two can be combined: the software is the foundation, and the automation handles the routine around it.
How to get started
The best starting point is a process that occurs frequently and is easy to check. In many businesses, that is either the journey from enquiry to draft quote, or incoming invoices. Either can be implemented as a single process.
- For one week, note how many enquiries and invoices come in and who handles them.
- Choose the process that ties up the most office time.
- Decide what may run automatically and what goes for approval.
- Test with real cases from recent weeks before the system goes live.
At the Scalient Group, a single process starts with the Starter package at €2,490 setup and €490 per month, excluding VAT. It is usually running in live operation after two to three weeks.
Conclusion
In the trades, the potential of AI lies not on the building site but in the office behind it. Pre-sorting enquiries, asking follow-up questions, matching documents and producing reports from voice notes: these are tasks that come up every day and lend themselves well to automation. What matters is that quotes, approvals and anything with real consequences stay with a person. That way the business gains time without giving up control.