AI knowledge base for car repair workshops: the practical guide

In a car repair workshop, an AI knowledge base makes findable what has ended up in heads, folders and inboxes over the years: solved special cases, manufacturers’ technical service bulletins, installation documentation and notes on regular customers’ vehicles. The workshop software remains the leading system for orders and invoices — the knowledge base answers the question nobody answers there: “Have we had a case like this before, and how did we solve it?”

In brief

• The most expensive hours in the workshop are diagnosis hours — and many of them arise because a case that was already solved cannot be found again.
• An AI knowledge base makes documented special cases, manufacturer information and conversion documentation findable by asking in plain language, with a source.
• A five-minute routine after every tricky case is enough to record the knowledge permanently.
• The workshop software knows what was done; the knowledge base knows how it was solved. The two systems complement each other.
• For safety-relevant values, the current manufacturer data remains the binding source.

Which cases cost the workshop the most time?

It is not the oil change and not the brake service that cost the workshop time, but the intermittent faults: the misfire that only occurs in the cold, the control unit that will no longer learn after a battery change, the rattling noise three mechanics have already searched for. In cases like these, experience decides between one hour and a whole day of diagnosis.

The problem: this experience is unevenly distributed. The foreman remembers that a similar case two years ago was down to a broken cable in the door grommet — but only if he happens to be there and remembers. If he is on holiday or long retired, the diagnosis starts from scratch. This very experience is a workshop’s most valuable knowledge, and it is documented almost nowhere.

On top of that comes the flood of documents from outside: manufacturers’ technical service bulletins and recall information, installation instructions from accessory suppliers, test reports, operating instructions for the workshop’s own equipment. Much of it is stored as a PDF somewhere — just not where it is needed for the specific vehicle on the lift.

What belongs in a car workshop’s knowledge base?

Everything that saves time the next time a comparable case comes up belongs in the knowledge base. For a typical independent workshop or franchised dealer workshop, this means above all five kinds of knowledge:

• Solved special cases: symptom, diagnosis, cause, solution — a short note with photos for each case, assigned to vehicle type and engine.
• Technical service bulletins and manufacturer information: filed as PDFs and therefore searchable by content, instead of being scattered across inboxes.
• Installation and retrofit documentation: tow bars, auxiliary heaters, gas systems, motorhome conversions — assigned to the customer vehicle concerned.
• Internal standards: from the process for the tyre storage service and the vehicle reception checklist to the procedure for the periodic roadworthiness inspection under § 57a of the Austrian Motor Vehicles Act, as the business organises it internally.
• Documents for workshop equipment: maintenance instructions and inspection certificates for lifts, the air-conditioning service unit, diagnostic testers.

What does not belong in it, on the other hand, are current order and invoice data — they stay in the workshop software — and the manufacturers’ up-to-date specifications, which are maintained in the respective manufacturer portal or data system.

What does the five-minute routine after a special case look like?

Recording knowledge only works in a workshop if it is quick. A fixed routine immediately after the case has been solved, while everything is still fresh, has proved its worth:

1. Note the symptom: two sentences, plus vehicle type, year of manufacture and engine — as the customer described the problem, and as it showed itself.
2. Record the diagnosis: what was checked, what could be ruled out? The dead ends in particular save the next person time.
3. Describe cause and solution: what was it really, and what was replaced, repaired or adjusted?
4. Attach photos: installation position, fault pattern, chafed spot on the cable — two or three mobile phone photos say more than a page of text.
5. Assign and upload: confirm vehicle type and topic, done. Automatic tagging suggests both.

It is important to keep the hurdle low: no mandatory forms, no neat handwriting, no approval loop. A rough but honest case note that gets found is worth more than perfect documentation that is never written.

What questions does the team ask day to day?

The benefit shows directly at the lift, when questions can be asked the way you would ask an experienced colleague:

• “Intermittent misfires on cylinder three with a cold engine — have we had a case like this on this model before?”
• “How did we handle the relearning procedure at the last gearbox replacement on this van type?”
• “What was retrofitted on the M. family’s motorhome, and where is the installation documentation?”
• “What does the technical service bulletin on coolant loss for this model series say?”

The answer comes in plain language and always names the source document — the case note from back then, the manufacturer bulletin, the installation documentation. One click opens the original. This is crucial, because AI can make mistakes: the mechanic checks the source before picking up a spanner. Systems such as junova, the AI knowledge base from Goma-it answer questions from your own filed documents and store them on servers in the EU, without customer documents being used to train third-party AI models.

Where do the workshop software and manufacturer portals fit in?

The knowledge base does not replace any of the existing specialist applications — it closes the gap between them. The workshop software remains the leading system for orders, times, parts and invoices: it knows what was done to a vehicle and what was charged. The repair history in the sense of accounting documents belongs there and stays there.

Manufacturer portals and technical data systems, in turn, are the binding source for current specifications: tightening torques, filling quantities, oil approvals, work instructions. Especially for safety-relevant values, the rule applies without exception: check the current manufacturer specifications before starting work, because filed documents can be out of date. The knowledge base helps here through transparency — every answer shows which document it comes from and how old it is.

Its own place is experience-based knowledge: the workshop software knows that three hours of diagnosis were charged two years ago. The knowledge base knows what came out of it.

How does a workshop start realistically?

The practical start consists of two steps. First: introduce the five-minute routine for every new special case straight away — this way the collection grows by itself. Second: document, with the experienced people, the ten cases that keep coming back, while this knowledge is still in the business. An afternoon with the foreman often achieves more here than months of collecting.

For the basic approach — clarify what the team searches for, define structure and access rights, fill the knowledge base, test with real questions — it is worth taking a look at the cross-industry guide AI knowledge base in your company. How the introduction works in concrete terms, from the initial filling to training the team, is shown in the overview of the introduction process.