AI knowledge base in your company: guides for trades, construction, law firms & more

An AI knowledge base in your company is a central, searchable collection of the company’s documented knowledge: instructions, service records, solved problems, tried-and-tested approaches and the experience that otherwise exists only in individual employees’ heads. It answers questions in plain language and always names its source. How much it helps depends heavily on the industry: a plumbing and heating business needs different knowledge to hand from a law firm or a manufacturer. This guide shows what knowledge and which everyday questions are typical of the main industry groups, where the line with the ERP system lies, and how to tell whether your company is ready to introduce one.

In brief

• The real value of an AI knowledge base lies in the company’s documented know-how: the “how” (settings, procedures, solved cases), not the “what” (records, orders, invoices).
• Records such as delivery notes and invoices belong in the ERP system; the knowledge base adds the experience that no ERP captures.
• Every industry group has its own typical kinds of knowledge and everyday questions – from the service record in the trades to the tried-and-tested wording in a law firm.
• A company is ready when knowledge depends on individual people, the same questions come up internally again and again, or knowledge is lost when staff change.
• The most common mistakes when introducing one are uploading old shared drives unfiltered, confusing records management with knowledge management, and not testing with questions from real day-to-day work.
• Answers must always come with a source; otherwise the system is unsuitable for business decisions.

Why is experience the most valuable untapped asset in your company?

Almost every small and medium-sized company has knowledge that is written down nowhere and yet needed every day. The service technician who knows that a certain heat pump always has to be re-paired after a power cut. The law firm employee who knows how a particular court deals with deadline extensions. The machine operator who runs the press with different parameters for thin sheet metal because otherwise the material warps. This knowledge has been built up at great cost over the years – and with every sick leave, every resignation and every retirement, a little of it leaves the company.

This is precisely the gap an AI knowledge base fills. It is not yet another store for records but the place where the company’s “how” is collected and made easy to find: how is the installation set up? How did we solve this problem last time? How does this procedure work here, and why do we do it this way? The difference from a traditional folder structure is that employees don’t need to know where something is or what the file is called. They ask a question in plain language and get an answer with a reference to the source document. The introductory article What is an AI knowledge base? explains in detail how such a knowledge base works technically.

The business case has two sides. First, the daily searching and asking around stops: instead of ringing the one experienced colleague who is out on site, the apprentice finds the answer himself. The article Time lost searching for documents shows where these constant interruptions come from. Second – and more important in the long run – the knowledge stays in the company even when people leave. A company whose know-how is documented and searchable is less vulnerable and easier to hand over; the Austrian Economic Chamber (WKO) also regularly highlights the risks of knowledge tied to individuals when it comes to business succession.

Where does the ERP system end and the knowledge base begin?

The most important distinction first, because it causes the most confusion in practice: an AI knowledge base doesn’t replace an ERP system and doesn’t try to. An ERP system (enterprise resource planning – the software for stock management, orders and commercial processes) manages business transactions: who ordered what, what was delivered, what was invoiced. Delivery notes, order confirmations, invoices and stock levels are records, and the ERP looks after them well.

The knowledge base answers a different kind of question. Not “Which invoice belongs to order 4711?” but “How did we solve the problem that came up with this order – and what do we need to watch out for next time?” The ERP knows that an installation was delivered; the knowledge base knows what settings it runs on, what faults it typically develops and what fixes them. The two systems complement each other: the ERP holds the company together commercially, the knowledge base holds its skills together.

In practice, this matters for what you upload: if you fill your knowledge base with thousands of delivery notes and copies of invoices, you dilute it. These records are already stored in a structured way in the ERP or the accounts and are quicker to find there. What belongs in the knowledge base is what the ERP can’t capture: the service record that goes with the job, the instructions with the technician’s handwritten additions, the email setting out the special solution agreed with the customer. The comparison Document management and AI explains how an AI knowledge base differs from traditional document management systems, which often sit between the ERP and the knowledge base.

What knowledge and which questions are typical of each industry group?

Five industry groups account for most Austrian small and medium-sized companies, and each follows its own pattern: what know-how does it produce, what form is it in, and what questions do employees actually ask day to day? The overview below shows typical cases – deliberately worded the way people really ask them in the workshop or the office.

Industry group	Typical documented knowledge	Typical everyday question
Trades and construction	Service records, installation settings, solved faults, assembly instructions with your own additions, photos of rating plates	“How did we set up the ventilation at the Bahnhofstraße property so that the whistling stopped?”
Law firms and consultancy	Tried-and-tested wording and templates, earlier assessments of similar cases, internal procedures, documented agreements	“What did we do last time when an objection deadline was missed?”
Retail and logistics	Solved complaints, know-how about dealing with suppliers, product quirks, internal process descriptions	“How do you make a complaint to supplier Huber – who is responsible there and what do they need from us?”
Manufacturing	Machine parameters, set-up instructions, documented fault fixes, inspection procedures, experience with materials	“What settings do we run the press on for thin sheet metal so that nothing warps?”
Public administration and associations	Resolutions with their reasoning, know-how about funding, internal guidelines, established procedures	“How did the claim for the regional funding go – what did we submit last year?”

Trades and construction: the knowledge is in heads, photos and scraps of paper

In trades and construction businesses, the gap between what people know and what is documented is at its widest. The technician knows what settings the heating system in the block of flats needs – but at best they are written on a note in his service case. The site foreman knows why the pipe was laid differently from the plan during the conversion three years ago – but nobody asks him until the next bit of damage occurs. Typical everyday questions sound like this: “What was the fix when the boiler at the inn kept going into fault F28?”, “What thickness of insulation did we put in for the loft conversion in Feldgasse?”, “How do you set the frequency converter on the old lifting station – the manufacturer no longer exists?”

The biggest gain here comes from combining easy capture with text recognition – OCR for short (optical character recognition, which automatically converts photographed or scanned text into searchable text). The photo of the rating plate, the photographed service record and the assembly instructions with handwritten additions become searchable without anyone having to type up reports in the evening. The key is that capture fits into the working day: a photo and two sentences straight after the job are worth more than the perfect report that never gets written.

Law firms and consultancy: wording and solved cases

Law firms, tax advisory firms and consultancies usually have their documents in digital form, but the firm’s real know-how is still hard to get at: it is spread across files, email inboxes and the experience of the longest-serving staff. The value lies less in finding individual documents again than in access to tried-and-tested solutions: “How did we word this contract clause last time, and why that way?”, “What was our line of argument in the similar case two years ago?”, “How does our deadline monitoring work when the colleague responsible is on holiday?”

For new employees in particular, this access is worth its weight in gold: instead of asking someone about every procedure, they can find the firm’s established procedures and model solutions themselves. At the same time, confidentiality and access rights matter more in this group than in any other – not every employee is allowed to see every file. For statutory texts and case law, people will carry on using the Austrian federal legal information system, RIS (ris.bka.gv.at); the knowledge base adds what is written down nowhere else: the firm’s own practice.

Retail and logistics: know-how rather than a flood of records

In retail, the instinct is to fill the knowledge base with records – after all, orders, delivery notes and invoices are produced in large numbers. That is precisely the wrong approach: these transactions belong in the stock management or ERP system. What is valuable is the knowledge around the transactions: “How do you handle a return with this supplier, and which photos do they want?”, “Which haulier takes bulky goods to Vorarlberg, and what do we need to bear in mind when booking?”, “Why do we always over-order this item – what was the reason?”

This kind of know-how usually sits with one or two people in the back office. If they are away, things grind to a halt. Documented solutions to complaints, supplier quirks and internal process descriptions make the company less dependent on individual people and shorten the time it takes new employees to settle in.

Manufacturing: parameters, set-up know-how and the machine whose manufacturer has gone

Manufacturers depend on experience that isn’t in any machine manual: the real parameters for a particular material, the order of steps when setting up, the trick for clearing the fault on the old machine whose manufacturer no longer exists. Typical questions: “How is the milling machine set up after a tool change – step by step?”, “What helped last time when the packaging machine kept drawing in the film?”, “What tolerances does the customer really accept for this series?”

A systematic approach pays off here: if every fault that gets fixed is documented in three sentences – symptom, cause, solution – a record of faults builds up over the months that measurably shortens downtime. Inspection reports and certificates complete the picture, but the core is the know-how of the people on the shop floor.

Public administration and associations: resolutions and the reasons behind them

In municipalities, federations and associations, the critical knowledge is often the “why behind the what”: a resolution is in the minutes, but the reasoning, the background and the alternatives considered at the time are known only to those who were there. Typical questions: “What has the board already decided on this, and why?”, “What conditions does the funding decision include, and how did we make the claim last time?”, “How does our procurement work – above what amount do we need competing quotes?”

Because terms of office end and volunteers come and go, loss of knowledge is built into the structure here. A searchable collection of minutes, guidelines and documented procedures makes every change of officers far less painful.

How can you tell whether your company is ready?

A company is ready for an AI knowledge base when important skills depend on individual people and the same questions keep coming up internally. You can establish both without outside consultants. Go through the following self-check honestly:

• There is at least one person whose absence means certain work doesn’t get done, because only they know how to do it.
• New employees take months to settle in because most things have to be explained to them in person.
• Experienced employees are interrupted several times a day with questions they have answered many times before.
• Problems that were solved once get solved from scratch a year later because nobody remembers the old solution.
• For properties, machines or cases from earlier years, every follow-up question starts with a lengthy search.
• Over the next few years, people with key knowledge will retire, and so far none of it has been recorded.
• Instructions and records exist, but they are scattered: network drive, email inboxes, paper folders, private notes.

If three or more of these apply, the pressure is usually high enough for an introduction to be worthwhile. If one or two apply, it is at least worth starting with a clearly defined area. The honest counter-check matters too: a two-person business where both people know everything and little needs documenting gets little out of it – and the same goes for a company that produces records in large numbers but has hardly any process know-how of its own. Records are the ERP’s job, and AI doesn’t change that.

What mistakes do companies make when introducing one?

Most failed introductions don’t fail because of the technology but because of avoidable mistakes in the preparation. The overview below shows the five most common mistakes, what they typically lead to and what to do instead.

Mistake	Typical consequence	Better approach
Uploading the whole network drive unfiltered	The AI quotes outdated instructions and superseded versions; within a week nobody trusts it	A deliberate initial upload: current, checked documents first, older material added selectively later
Uploading records instead of knowledge	The knowledge base becomes a poor ERP, and the real knowledge questions go unanswered	Leave delivery notes and invoices in the ERP; upload instructions, records and solved cases
Sorting out access rights only after uploading	Confidential cost calculations or files are visible to everyone, and the project is halted	Set up users and roles before uploading, and add sensitive areas last
No test questions from real day-to-day work	The system does well in the demo but fails the technician’s first serious question	Before going live, collect twenty real everyday questions from employees and work through them
Introduction without anyone in charge	Once the initial enthusiasm wears off, nothing new is added and the knowledge base goes out of date	Name one person to check new additions and remind the team to document

A sixth mistake deserves a paragraph of its own: expecting the AI to know things that aren’t documented anywhere. A document-based AI answers only from the documents it has; it cannot supply what was never recorded. That is a strength, because – unlike chatbots trained on anything and everything – it prevents made-up answers. But it also means that introducing one shows up, mercilessly, where documentation is missing in the company. Companies that treat this as a task rather than a flaw in the system get the most out of it.

The last point is data protection. The GDPR (General Data Protection Regulation, the EU rules on handling personal data) requires, among other things, a data processing agreement (DPA) with the provider and clarity about where the data is stored; the Austrian Data Protection Authority provides guidance. Servers in the EU and a guarantee that company documents won’t be used to train AI models should be minimum requirements. The guide AI and the GDPR in Austria sets out what this means in practice for Austrian companies.

How does knowledge get out of people’s heads and into the knowledge base?

The most common objection to a knowledge base is: “Nothing’s written down here anyway.” That is hardly ever entirely true – and where it is, it is not an obstacle but the very reason for the project. You don’t get knowledge out of people’s heads with a one-off effort, but with small habits linked to specific occasions.

The three-sentence rule after every solved problem works well: what was the problem? What helped? What should we watch out for next time? These three sentences – typed, dictated or as a photo of a note – are enough, because the AI takes care of finding them again. Nobody has to maintain folder structures or assign keywords; the system’s automatic tagging handles that.

For the initial upload, this checklist works well:

• Collect the ten most common internal questions – they show which knowledge needs to go in first.
• Go through existing instructions, records and templates and mark the current versions.
• Allow one to two hours per key person to note down the most important topics they carry in their heads as keywords.
• Collect photos of rating plates, installation sheets and handwritten notes and capture them with text recognition.
• Deliberately weed out outdated and duplicate documents rather than carrying them over.
• Decide who will record new solutions in future, and when (for example straight after the service call).

Where existing knowledge already sits in other systems – in email inboxes, in Microsoft 365, on the network drive – there is no need to transfer it by hand: the article Connecting documents and systems describes how existing shared drives and programs are connected.

How does the introduction work, whatever the industry?

Introducing an AI knowledge base follows the same five steps in every company; only the content varies by industry. A tried-and-tested process looks like this:

1. Intro call: establish which questions keep coming up in the company and where the knowledge to answer them sits today – in documents, systems or people’s heads.
2. Analysis: map out the kinds of knowledge, where they are stored and who has access; decide what goes into the initial upload and what deliberately stays out (such as records that stay in the ERP).
3. Set-up and initial upload: set up the system, create users and roles, upload the selected documents and record the first topics from people’s heads.
4. Test questions: use real questions from day-to-day work to check that the answers are correct and the sources match.
5. Introduction for the team: show employees how to ask questions and record knowledge, so that the system gets used every day and keeps growing.

The test-question step is the most important and the one most often skipped. The system is only ready for everyday use once the office manager, the technician or the law firm employee have asked their own typical questions and got useful answers backed by sources. The How it works page describes the process in detail.

A small or medium-sized company doesn’t need its own IT department to do this. Systems such as junova, the AI knowledge base from Goma-it are designed so that set-up and the initial upload happen with support, after which the company carries on without specialist technical knowledge. The practical report AI knowledge base in practice shows what this looks like in real companies – with typical questions, the initial upload and everyday use.

What can an AI knowledge base do – and what can’t it do?

An AI knowledge base finds and explains documented knowledge, but it replaces neither industry-specific software nor professional judgement. Being honest about where that line lies saves disappointment and false expectations in the team.

The tool is well suited to anything that exists as text or can be turned into text with text recognition: instructions, records, solved cases, templates, emails setting out agreements, photographed notes and rating plates. It is no substitute for ERP and accounting software, costing programs or CAD systems (computer-aided design – programs for technical drawings and plans); their data stays in that software. The knowledge around these systems does belong in it, though – such as the internal instructions on how to create a credit note correctly in the ERP, or the description of the plans for the CAD project.

Just as important: the AI’s answers are a basis for work, not a decision. In a law firm, the lawyer still checks the legal position; in a plumbing and heating business, the master craftsman still decides on the technical solution. The knowledge base shortens the route to the relevant information – responsibility for applying it stays with people, and every answer must be quick to check against its source.

What is the best way to get started in your industry?

The best start is a small, clearly defined area where the lack of knowledge hurts most today. A plumbing and heating business might start with the service records and installation settings for existing customers, a law firm with the templates and documented procedures of recent years, a manufacturer with the set-up instructions and fault records for one group of machines. What matters is that the starting area covers real everyday questions – the winner isn’t the area with the most files but the one that generates the most follow-up questions.

The starting area teaches you two things: which questions your employees actually ask, and how good the documented knowledge needs to be for the answers to be useful. With that insight, you can expand the knowledge base step by step without taking on a major project. And with every case that is recorded, the company becomes a little less dependent on what individual people know.

More detailed guides to individual industries – for trades businesses or construction companies, for example – will follow in this Guides section, covering their particular kinds of knowledge, example questions and limitations in detail.