Electrical and systems engineering · South Tyrol
Digital assistant for a trades business: materials, extra work and technical questions by voice
Part of: AI automation
Also relevant for: ERP & integration
The starting situation
Installers interrupt their work to record materials in several apps. Reports on extra work and service completions consist of many individual steps, and answers to technical questions are scattered across manuals, data sheets and project files.
How we solve it
Alpino AI builds an assistant for the smartphone: employees speak, take a photo or type. It queries stock levels in the ERP, prepares material requests and work reports as drafts and answers technical questions with a source.
What you get
A few steps instead of many entries, with bookings that reach the ERP automatically once confirmed. The assistant suggests, the person confirms, and every action is traceable.
Do you recognise a similar process in your company?
If this use case feels close to your daily work, we can assess data availability, effort and a realistic first MVP together.
Why AI projects work →Starting point
In a trades business with several sites, questions and bookings arise every day and partly end up in different apps: material requests in one app, hours in a second, service jobs in a third. An example from the company: an installer interrupts the installation several times just to record material on a phone. Completing a service job then involves technician, type of job, duration, note, signatures, follow-up activity and PDF dispatch as individual steps.
Knowledge is another issue: data sheets and manuals sit on the server, wiring diagrams in the project files, project data in the cloud. Anyone on site who needs to know something calls the office.
The solution
Employees, from the construction site to the office, ask questions and give instructions by voice, photo or text on their own smartphone, Android or iPhone. The assistant is a dedicated app, so that follow-up questions, overviews and approvals can be shown clearly. The use cases, with the first two as the main ones:
- Stock query and material request: one voice message with several items is enough. The assistant shows an overview of articles and quantities, the installer confirms, the request goes to the ERP.
- Extra work and additional orders: dictating instead of filling in forms. The report is created as a draft, checked and approved. Completing a service job is part of this.
- Technical questions: answers from manuals, data sheets and wiring diagrams, always with a source.
- Reading plans: from plans in PDF format, a proposal for the bill of materials and the list of control elements is created, which the responsible person approves.
Process and pilot
All conceivable use cases and roles at once would be too broad a start. The start therefore covers the two main cases, a small pilot team and a clear task: few steps, simple input. Speech should be tested early, since employees speak dialect and partly Italian. Only when material and extra work run reliably are further cases added, such as goods receipt via a photo of the delivery note or invoice checks.
What stays with people
The principle is: AI suggests, a person confirms. Figures come exclusively from the ERP, never from the model's memory. Technical answers come only with a source. For safety-relevant topics such as troubleshooting protective devices, the assistant only outputs the test instructions stored by the company. Bookings are created only after confirmation; checking the material request in the office remains a person's task.
Connection to existing systems
The ERP is the central system. The assistant reads stock levels, articles and orders and writes requests and reports back through approved, logged actions. Manual re-entry is not necessary; exactly how writing works depends on what the ERP manufacturer's interface offers and is clarified at the start. In addition there are the file server with manuals, data sheets and project files, and the cloud storage for project data. Technically the assistant is a chat with tool calls: it can only trigger approved actions and has no free access to the systems. A related case is described in Knowledge assistant on the file server, with background in the article AI agents for business processes.
Transferable to other companies
The pattern suits companies with field service, construction sites or service work, where knowledge and bookings are far apart: building services, construction, plant maintenance, facility services. More on the service on our AI automation page and on the ERP connection on our ERP integration page.
If you would like to check which processes in your company suit an assistant, book a 30-minute initial conversation.
This use case is based on a real client project. Sector and region are named, the company itself remains anonymous.