AI in the Skilled Trades: 5 Realistic Use Cases for Businesses in the Ruhr Region

When artificial intelligence comes up in the skilled trades, the conversation usually turns to robots on site or machines that plan themselves. The real benefit today lies elsewhere: in the office, in the van, in the evening that starts later because three quotes still need writing. That is exactly where the use cases that already work today apply – including in a five-person business in Recklinghausen.

Note: This article deliberately avoids product recommendations and numerical promises of success. Where reliable survey data is missing, the benefit is described in qualitative terms. Status: August 2026.

How widespread is AI in the skilled trades really?

Only four percent of German skilled trade businesses currently use artificial intelligence. That is the finding of the study "Digitalisierung des Handwerks" by the German digital industry association Bitkom, conducted by Bitkom Research. The study surveyed 504 trade businesses by telephone; fieldwork took place in June and July 2025 and the report was published in early 2026.

That figure sits well below the impression created by trade fairs and vendor marketing. For businesses in the district of Recklinghausen this means two things. The advantage gained from a single, properly introduced use case is real – and the pressure to change everything at once is not.

Source: Bitkom – Digitalisierung des Handwerks (study report)

Which use cases already work in day-to-day operations?

Five areas deliver the fastest return in small businesses because they attach to work that already exists and require no new hardware. All five can be implemented with mainstream language models and the office software you are already running.

1. Drafting quotes, letters and customer correspondence

The classic case, and the one with the lowest barrier to entry. You enter the key facts as bullet points – trade, scope of work, materials, scheduling – and have a properly worded covering text for the quote generated from them. The same applies to declining a job, rescheduling an appointment, escalating a payment reminder or responding to an irritated customer email.

The AI does not do your costing. Prices, measurements and labour estimates remain your responsibility. What disappears is the struggle over wording at the kitchen table at 8 p.m. Sole traders and small firms in particular report noticeable relief here, because correcting a draft is faster than writing from a blank page.

2. Triaging email and preparing reply drafts

Enquiries, supplier queries, appointment confirmations and advertising all land in the same inbox. Current AI features in Microsoft 365 and Google Workspace can sort incoming mail into categories, summarise long threads and prepare a reply draft that you only need to review and send.

The benefit lies less in the typing than in the sorting. Anyone who can see in the morning which three of 40 emails genuinely need answering today loses less time to context switching. If you are weighing up Microsoft tooling anyway, the article on Microsoft 365 Copilot for SMBs helps put the options in perspective.

3. Organising photo and site documentation

Every job site produces photos – before the work, after the work, of the damage, of the defect. The problem is rarely taking them; it is finding them again. AI-supported image recognition can group shots automatically by project, room or trade and generate descriptions that remain searchable later.

Voice input works well alongside it: instead of typing up notes in the evening, fitters dictate a short report on site, from which a structured text with date, property and open items is produced. For businesses dealing with warranty claims, that is the difference between "somewhere on my phone" and documentation that holds up.

4. Easing scheduling and dispatching

Anyone working across the Ruhr region between Recklinghausen, Herten, Marl and Gelsenkirchen loses time mainly to poorly sequenced routes. AI features in trade management software and calendar systems can propose appointment slots based on travel distance, urgency and staff qualifications instead of leaving you to piece the week together by hand.

The realistic view: the dispatcher decides, not the system. Short-notice emergencies, customer preferences and the state of the team appear in no database. The gain is a solid first draft rather than an empty weekly plan.

5. Making your own documents searchable

Data sheets, installation instructions, manufacturer approvals, old bills of quantities: in many businesses this knowledge sits in folders on a server or in the head of the longest-serving journeyman. AI systems can draw on that body of material and answer questions in plain language – including a pointer to the document the answer came from.

This is technically more demanding than the first four cases and pays off mainly where a large body of documentation exists. The article on vector databases and RAG describes how such a setup is built.

Where does data protection become critical?

As soon as customer names, addresses, photos taken inside private homes or invoicing data are entered into an AI tool, you are processing personal data – and you need a contractual basis for it. In the trades this affects almost every use case listed above, because a quote without customer data does not exist.

A private free account is not sufficient. Which plan makes a data processing agreement possible, and what else to watch out for, is covered in detail in the article on using ChatGPT in a GDPR-compliant way. On top of that comes the AI literacy obligation under Article 4 of the EU AI Act: anyone deploying AI must ensure that the staff involved have a sufficient level of AI literacy.

Where should a business in the Ruhr region start?

With exactly one use case covering a task that comes up every week and is a nuisance today. A realistic entry looks like this:

1. Pick the task: Name the activity that costs you the most evenings – usually quote texts or documentation.
2. Settle data protection before tool selection: Define which data may be entered first, then choose the provider.
3. Run a four-week trial: One employee, one use case, then an honest review. If nobody misses the tool, it was the wrong case.
4. Write down the rules: One page is enough – permitted data, off-limits data, point of contact. Together with a short briefing for the team, this covers the AI literacy obligation under Article 4.

Neutral local starting points are the Handwerkskammer Münster, whose chamber district covers the district of Recklinghausen, and the Kreishandwerkerschaft Recklinghausen. What external support costs is set out in the article What does AI consulting cost for small businesses?.

Conclusion

AI in the skilled trades is currently less a question of technology than of prioritisation. The five use cases described here attach to work that exists in every business, and none of them require investment in machinery. Start with one case, settle data protection beforehand and take an honest look after four weeks: the risk is negligible – and in the district of Recklinghausen you will be joining a still very small group.

Would you like to know which of these use cases fits your business and what the first step looks like in practice? For trade businesses in Recklinghausen and the wider Ruhr region I am happy to take a look – straightforward, jargon-free and without obligation: AI consulting for small businesses.

Publication note: This article was scheduled for 28 July 2026. Because of a technical fault in our publishing automation, it did not go live until 11 August 2026. All information was re-checked for accuracy before publication.