AI knowledge base for trades businesses: a practical guide

junova editorial team · 13 July 2026

An AI knowledge base for trades businesses is a central store for quotes, instructions, delivery notes and site photos that anyone can search by asking a question in plain language. Instead of rummaging through folders or ringing the office, the technician asks “What were the settings on the system in Bahnhofstraße?” and gets the answer together with the source document. For businesses with piles of paperwork and a full diary, it is an obvious first step into AI, because it solves a real everyday problem.

In brief

Trades businesses lose time because knowledge is scattered across paper, emails, phones and people’s heads.
Text recognition (OCR) makes even scanned delivery notes, old instructions and photographed notes searchable.
Technicians upload photos and documents straight from their smartphone, with no special software to install.
Roles keep what the office sees (cost calculations, personnel matters) cleanly separate from what the installation team needs (instructions, job details).
Getting started is not an IT project: an analysis, an initial upload of the most important documents and some test questions are all it takes.

Why are trades businesses drowning in paperwork and scattered files?

Trades businesses create documents in lots of places at once, but rarely is there one place where everything comes together. The quote is written in the office, the delivery note turns up on site, the photo of the finished connection ends up on the technician’s private phone, and the special arrangement with the customer is buried in an email from the boss. Each of these works fine on its own; put together, they add up to paper chaos.

A plumbing and heating business with 15 employees soon has thousands of relevant documents: years of quotes and order confirmations, maintenance instructions for the equipment it has installed, inspection reports, time-and-materials reports, photos of pipe runs before plastering. Only a tiny fraction of this is ever needed – but that is precisely the part nobody can find when it matters.

The consequences are plain to see day to day. The technician rings the office and waits, the office worker stops what she is doing to search, and if she finds nothing, someone drives out to the site or the job gets done from memory. A searchable central store tackles exactly this problem.

Which questions does an AI knowledge base answer in day-to-day trades work?

An AI knowledge base answers the questions that currently end up as a phone call to the office or to the most experienced colleague. Typical examples from plumbers, electricians, joiners and similar businesses:

“What did we quote the Muster family last year, and what was included in the price?”
“What were the heat pump settings at the Gartenweg property after the last service?”
“Which make and model did we install for this customer?”
“Are there photos of the pipe run in the bathroom before it was plastered?”
“What does the manufacturer’s manual say about error code E14 on this boiler?”
“What did we agree with the customer about the warranty on the custom-made item?”

The key difference from a traditional folder search: the answer comes back as text with its source, and one click opens the original document. The technician doesn’t need to know what the file is called, which folder it’s in or who saved it.

To set expectations: the AI bases its answers on the company’s own documents. No system in the world can conjure up something that was never written down, such as a verbal promise. In practice, this often prompts businesses to put important agreements in writing more consistently, because the effort now visibly pays off.

What does text recognition (OCR) do for the paper archive?

OCR stands for optical character recognition: software recognises the text in scans and photos and makes it searchable. For trades businesses this is a key building block, because so much of their knowledge exists only on paper: delivery notes with handwritten comments, old instructions, signed acceptance reports, fax confirmations from years gone by.

With text recognition, a photographed delivery note becomes a searchable document. Ask “When was the material for the Hauptplatz site delivered?” and the search will also find the delivery note that was uploaded as a lopsided phone photo. Older PDF files that are stored only as images become searchable in the same way.

Be realistic about the quality of the originals, though. A clean scan works very well, a crumpled carbon copy with faint writing defeats any text recognition, and hurried handwriting remains a weak spot. The practical rule: photograph important papers as flat as possible and in good light, and for critical handwritten notes, type up the key point briefly as well.

How do documents and photos get from the site into the knowledge base?

Technicians upload directly from their smartphone in the browser – nothing special has to be installed on their phone. The technician photographs the rating plate, the meter reading or the finished connection and uploads the photo. From then on, the information is available to the office and for future jobs.

Automatic tagging does most of the filing. The system reads each uploaded document and fills in details such as customer, project, installation, date and topic by itself. A person only confirms the document type in the review area, because it decides who will be able to see the document later; for a straightforward delivery note that is one click. Nobody has to sort documents into folder structures by hand any more – and that is exactly where traditional filing systems in the trades keep failing: they only work as long as someone keeps the folders in order with real discipline.

It also solves the problem of photos on private phones. Site photos that currently vanish into private galleries and messaging chats end up somewhere that belongs to the business, even if the employee moves on one day.

Who sees what? Roles for the office and the installation team

User and role management decides which employees can see and search which documents. In trades businesses, a simple two-way split works well and can be refined as needed.

Role	Typically sees	Typically does not see
Office and management	All documents including cost calculations, purchasing terms, personnel documents	–
Installation and site	Instructions, job details, reports, photos, technical documents	Cost calculations, purchase prices, personnel files
Apprentices	Instructions and work instructions	Customer correspondence, commercial documents

Define the roles before the initial upload, not afterwards. If you upload everything first and restrict access later, sensitive documents may have been visible to everyone in the meantime. A quick think about “Who needs what to do their job?” is usually enough.

Security also means looking outwards: servers in the EU, a data processing agreement (DPA) under the GDPR (General Data Protection Regulation, the EU data protection law) and a guarantee that your documents won’t be used to train AI models are essential for a trades business’s customer data, not an optional extra.

How does the introduction work without a dedicated IT project?

A trades business can introduce an AI knowledge base within a few weeks, without servers in the basement or an IT department. The process follows the cross-industry pattern described in the overview article AI knowledge base in your company: intro call, analysis, set-up with an initial upload, test questions and an introduction for the team.

In the trades, that means starting with the documents that are most useful – usually quotes and job documents from the last few years, instructions for the installations under ongoing maintenance and the reports for active customers. There is no need to digitise the entire paper archive back to the day the business was founded; further scanning follows what day-to-day work actually requires.

The real test is a set of questions from everyday work. When the office manager asks her five most common search questions and the most experienced technician asks his typical on-site questions, it soon becomes clear whether the upload and the answers are right. Providers such as Goma-it from Vorarlberg structure the whole introduction along these lines; the How it works page describes the individual steps.

Once it is up and running, the system largely looks after itself, because new documents are added continuously as they are uploaded. One task remains: confirming the document type in the review area so that the access rights are correct. It therefore makes sense for one person in the office to act as the contact and handle these approvals alongside their other work; in a typical business, there is no other ongoing maintenance.

Frequently asked questions

What does searching for documents really cost a trades business?

The cost lies in interrupted work: the technician waits for the office to ring back, someone in the office leafs through folders, and the boss spends the evening answering questions a filing system could have answered. On top of that come mistakes when people work from memory because they can’t find the documents. How much this adds up to varies from business to business, but it is noticeable almost everywhere.

Does an AI knowledge base also work with handwritten notes?

Handwriting is the hardest test for text recognition: clear block capitals usually work, hurried joined-up writing often doesn’t. In practice, a good approach is to photograph important handwritten notes and also type up the key information briefly. That way the original is kept and the information can be found reliably.

Do technicians need training for this?

Very little, because using it matches what technicians already do every day: take photos and type or dictate a question. A short introduction of under an hour is usually enough. What matters more than training is that the first questions people ask get useful answers – then the team starts using the system of its own accord.

Who looks after the knowledge base day to day?

Day to day, the knowledge base largely keeps itself in order: when documents are uploaded, the system fills in details such as customer, project, installation, date and topic by itself. Only the document type needs to be confirmed, because it decides who can see the document; for straightforward documents that takes a few clicks in the review area. It makes sense for one person in the office to handle these confirmations and flag outdated documents. A fixed weekly maintenance day isn’t necessary.

What happens to the old paper archive in the basement?

There is no need to digitise the whole paper archive; for most businesses that would be out of all proportion. What works well is to scan only what is still needed: documents for current maintenance customers, instructions for installations still in use, quotes for active customers. Everything else stays in the basement and is scanned only if and when it is needed.

How secure is customer data in an AI knowledge base?

What matters is where the servers are, what the contract says and who has access. Look for servers in the EU, a data processing agreement (DPA) under the GDPR and a guarantee that your documents won’t be used to train AI models. Within the company, roles decide who can see which documents.

For your company

To see what this looks like in practice, visit the page for your industry – with the questions people there ask and the documents the answers come from.