Skip to content

AI for electrical, plumbing and heating businesses: service calls, advice, maintenance

By the KI können editorial team7 min read

Written with AI and checked against the sources by a second AI. As of November 16, 2026. More in the notes on AI.

In an electrical or plumbing and heating business, the phone rings a lot: the heating is off, a breaker keeps tripping, questions about heat pumps. Then there are maintenance visits, test reports and quotes. An AI can write you call guides, customer information and templates. Readings, standards and safety questions stay with you and the manufacturers’ documents.

An electrician stands with a tablet in front of the open meter cabinet, a ladder, a cable drum and a toolbox nearby.

Key points

  • A guide for taking fault reports helps the office ask the right questions. The AI drafts it, you check it.
  • The AI writes customer information on heat pumps, solar panels or wall boxes from your facts. Look up funding and laws with the funding body.
  • Readings, test results and settings come from your measurement and the manufacturer’s documents, never from the chat.
  • An AI can sort fault codes into categories but can’t reliably interpret them. What counts is the device’s manual.
  • Addresses, meter numbers and photos showing house numbers don’t belong in a private AI account.

Which tasks in the business suit AI

An AI like ChatGPT, Claude, Gemini or Copilot is good at phrasing, organizing and explaining. It has never seen the system at the customer’s home and knows the manufacturer’s specifications only roughly. What it needs to know goes into your instruction.

Tasks in an electrical or plumbing and heating business at a glance
TaskYou provideYou check
Guide for taking fault reportsHow you ask and decide todayWhether an emergency is correctly recognized
Customer information on heat pumps or wall boxesYour facts, your proceduresStatements on funding, law, technology
Maintenance reminder and checklistManufacturer documents and your procedureWhether a test step is missing
Letter on a test resultResult and defects from your reportEvery detail against the report
Explaining quote optionsYour cost calculationEvery figure

For descriptions of work, measurement notes and materials lists, there is a separate guide: ChatGPT for tradespeople.

In the trial lesson you put your first instruction together and get feedback. No account, in a few minutes. Start the trial lesson

Taking fault reports: a guide for the office

Aisha runs the office of a plumbing and heating business (SHK, short for Sanitär, Heizung, Klima) with twelve people. In winter, customers whose heating isn’t working call every day. So that the technician brings the right things, she needs the same details on every call. She has the AI build a guide from her notes:

Sample instruction
Turn my notes into a guide for taking fault reports in a plumbing and heating business.
For three cases: heating off, no hot water, water leaking.
For each case at most six questions in everyday language that customers can answer on the phone, for example: does the device show a number or a symbol?
At the end of each case: which details do I write down for the technician?
Don’t write any rules on emergencies and no do-it-yourself tips. I’ll add those from our emergency instructions.
My notes:
[paste the notes here]

The master craftsman (Meister) reads the guide before it goes into the office. What to do if there is a smell of gas, water in the fuse box or an electrical accident is set out in your company’s emergency instructions, and for gas also in the gas network operator’s guidance. Take these sentences from there, not from the AI.

Customer information on heat pumps, solar panels and wall boxes

Many customers ask the same questions: does a heat pump suit my house? What do I need for a wall box? How does it work with the grid operator? Femi is a master electrician and wrote an information sheet on this, with the AI and his facts:

Sample instruction
Write an information sheet for private customers: “How we install a wall box”.
At most 250 words, formal, no jargon, in steps.
Our steps: on-site appointment to look at the meter cabinet and cable route, quote, registration with the grid operator, installation, testing and instruction.
Mention: wall boxes must be registered with the grid operator before they are put into operation. If the combined power of all charging points at the house connection exceeds 12 kVA (kilovolt-amperes), the grid operator must agree beforehand. A typical 11 kW wall box on its own stays below that. We take care of this for you.
No statements on funding, costs or charging times.

Femi looked up the rule on registration himself in the Low Voltage Connection Ordinance (Niederspannungsanschlussverordnung) and wrote it into the instruction. That way the AI can’t build in anything outdated. This is especially important with funding programs and the rules on replacing heating systems: they have changed several times in recent years. What currently applies is stated by the funding body and in the text of the law. Many AIs only know an older version.

Maintenance, reminders and checklists

Emil is a plumbing and heating technician and looks after the company’s maintenance customers. He has the AI prepare three things: a reminder of the annual maintenance, a short text on what happens during maintenance, and a checklist for the appointment. He builds the checklist from the manufacturer’s maintenance documents, which he gives the AI as text. He takes test steps and settings only from there.

Sample instruction
Turn the “Maintenance” section of the manufacturer’s manual below into a checklist to tick off at the on-site appointment.
Take every test step and every value exactly as it appears there. Don’t add anything and don’t leave anything out.
Format: number, test step, target value from the manual, field for the reading, field for remarks.
[paste the section from the manual here]

Then Emil puts the checklist next to the manual and counts the test steps. If one is missing or a value is different, the list is no use.

After the appointment

Many customers want to know what was done during maintenance. After the appointment, Emil dictates three or four sentences into his phone: what was checked, what was replaced, what is due next time. The AI turns this into a short report in everyday language, without values that aren’t in the note. If the report recommends a repair, Emil checks whether he really said that. The customer reads a recommendation in the report as a promise.

Test reports: values from your measurement

For tests of electrical installations and devices there are reports, often straight from the measuring device’s software. Here an AI shouldn’t add any values, name any limits or assess any result. What it does well: a letter to the customer that explains the result in plain words.

  • Give it only what is in the report: passed or not, which defects, to be fixed by when.
  • Leave out name, address and installation ID and insert them afterwards.
  • Check every defect in the letter against the report. A wrongly described defect can become expensive later.

Fault codes and technical questions

Ask an AI what a fault code on a heating unit means and it often comes up with a plausible answer. Whether it fits exactly this device and this model series, you don’t know. Use it at most to sort out what to look up in the manual or with the manufacturer’s hotline.

The same goes for standards and regulations, cable cross-sections, circuit protection, flue gas, gas and refrigerants. What counts here is the standard in the original, the manufacturer’s documents, the grid operator’s requirements and your measurement.

Comparing two quotes in plain words

Customers find it hard to compare two options, such as a new gas boiler and a heat pump, or a small and a large solar power system. From your cost calculation, the AI can make a side-by-side comparison in everyday language: what is included, what the customer has to do themselves, what changes in operation. You provide the figures, and you tell it not to add any savings, funding amounts or payback periods. If it calculates something itself, recalculate at least one line.

Customer data and photos

Names, addresses, phone numbers and meter numbers don’t belong in a private AI account. Be careful with photos too: a photo of the meter cabinet often shows the meter number, one of the front door the house number, one of the boiler room perhaps mail on the shelf. Crop out such details or don’t use the photo at all. How business accounts and data processing agreements (DPA) help is explained in the guide Data protection and AI tools.

Building call guides, information sheets and checklists with an AI and testing them on a case you know is what you practice in the learning app “KI können”, in short lessons with examples from the trades. You do the actual work in your own AI. The trial lesson takes no account.

Sources

Information as of November 16, 2026. Laws and offers can change; this article is not legal advice.