AI for trades businesses: automating the office work
AI for trades businesses: automate the office work first
AI for trades businesses pays off where someone is still typing things up in the evening: in the office between the site and your trade software.
You can automate most of the office work in between: moving day reports across, ordering material, filing photos against the right job, preparing invoice drafts and flagging payments that are due. Costing, prices and promises to customers stay with you. The building site and the website change little.
This piece is for owners and office managers of Swiss trades and construction businesses, roughly 5 to 50 people, working with trade software such as Sorba, Messerli or another package. It shows where office work gets stuck, why the usual fixes often leave a gap, and what a tighter workflow looks like.
Where the office work gets stuck
The pattern is double entry: captured outside, written again in the office. The trade software knows the job. What happens on site arrives on paper, as a photo in a group chat, or by phone. Many firms search for «automate daywork report» or «typing up site reports» even when the bottleneck is wider than one form.
| Task | How it often runs | What can run automatically | What stays with you |
|---|---|---|---|
| Quotes | Copy the last quote, adjust the lines | Draft from a template and the customer record | Costing and price |
| Daywork and day reports | Paper in the van, typed up on Friday | Capture on the phone or a photo of the paper, fields pre-filled | Checking what gets billed |
| Material orders | A call or message to each supplier | One order draft per supplier from a single list | Approving the order |
| Scheduling | Whiteboard or Excel, changes by phone | Weekly plan from the jobs, sent to the crew | Who goes where |
| Invoices | Hunt down the reports, type in the lines | Invoice draft from checked reports | Approval and sending |
| Payment reminders | Go through open items when there is time | List of overdue items and a draft reminder | Whether and how to chase |
| Photo records | Photos on private phones and in chats | Filed against the right job, with the date | Which photos go to the customer |
Usually it is two or three rows, and one of them lands on the same person every week, late in the evening.
Why the trade software alone does not close the gap
Trade software works with what is in it. The gap is the way in. Many packages have their own apps. Sorba, for example, offers a day report app that records staff, machines and material on site and sends the report to the office. If every crew uses an app like that every day, this part is often done.
In many businesses not everyone does. The subcontractor brings paper. The fitter who has been with you for twenty years prefers a pen. Material from the DIY store sits on a receipt in the glovebox. Extra work was agreed verbally on site. Photos live in a group chat on private phones. The software sees none of it until someone in the office types it in.
A ChatGPT subscription is the second usual answer. According to the AXA SME labour market study, cited by DeepCloud, Swiss SMEs use AI most often for translation (52%) and correspondence (47%). That helps with a letter to the property manager. It does not get the report from the van into the invoice.
The third answer is another person in the office. That can be the right call. In smaller businesses, though, the owner often does the office work in the evening because the site takes the day.
What a tighter office workflow looks like
The office stays. Each piece of information is captured once, where it happens. The office checks drafts instead of retyping.
Today
- Reports on paper, typed up on Friday
- Photos in a chat, missing from the job file
- The invoice goes out when someone has time
- Open items get noticed when the account runs low
Tighter workflow
- Report on the phone or as a photo, fields pre-filled
- Photos sit with the job, dated
- The invoice draft waits for approval
- Every Monday, a list of overdue items
AI has a small, clear place in this. It reads a photographed paper report, matches «2 bags of cement» to the right item in your list, and flags what it cannot read with confidence. Ordinary automation does the rest: moving data to the right place, producing a draft, setting a reminder.
Material orders follow the same pattern as KitchenCrew, a product we built for restaurants: the team types a list into a chat, and an order draft is created for each supplier, which someone checks and sends. A kitchen is not a building site. Copying one list out five times for five suppliers is the same job.
What applies to photos and customer data
Site photos often show more than the work: flats, names on doors, plans. Images like that do not belong in a free personal account of an AI service. What to watch for in Switzerland is covered in ChatGPT and customer data. Filing works the same as for paper: with the job, not on the fitter’s phone.
Who should not start yet
A three-person business writing five invoices a month usually does not need automation. It needs a clean template and a fixed slot in the diary. The same goes if every report looks different and nobody has decided which fields it must have. Then a single standard form comes first, automation after.
If your crews already use the app of your trade software every day, the manual work is somewhere else, perhaps in payment reminders or scheduling. Which workflows are often worth it in an SME office, trades or not, is in three workflows worth automating.
The first step
Take last week’s reports and count: how many there were, who typed them up, how long it took, and how many days passed between the work and the invoice. Those four numbers show whether automation pays. Use the four questions in which process to automate first to sort the other candidates. How the path from daywork report to invoice works in detail is the subject of a separate article.
If you want to look at one specific task from your office, the Process Check is made for that: 30 minutes, one task, and afterwards you know whether it can be automated and what that would cost.
Written by
Aurum Avis Labs
Builds and ships at Aurum Avis Labs. Writes here about what we learn working with founders and SMEs in the DACH region.
Related Articles
You might also be interested in these articles
From daywork report to invoice: five steps for trades offices
From daywork report to invoice in five steps: capture, structure, check, invoice draft, approval. With an example and when the trade app is enough.
How to follow up on quotes without losing them
How to follow up on quotes so they do not die in the inbox. Set a cadence, one owner, and a CRM status. Automation may remind and draft.
Which process should an SME automate first?
Which process should an SME automate first? Four questions on frequency, pain, a clear end and data. The check is free.