AI for small businesses: what works today and where the limits are
Where AI helps small trade and service businesses in everyday work, what to clarify first about control and data protection, and how to start small.

A few years ago, artificial intelligence was mostly something for large corporations with their own data teams. Today, a painting business can have AI draft a quote text from bullet points, an electrician can answer a customer email in Turkish without speaking Turkish, and a cleaning company can have supplier invoices read automatically.
The difference: AI now understands everyday language. You don't need to know how to program to use it. Many tools are available as a monthly subscription.
This article shows where AI can help in the everyday work of small businesses, what you should clarify first and how to start small. Without promising that everything will then take care of itself.
In short
- AI mainly helps with texts, enquiries and documents: writing drafts, sorting requests, reading receipts.
- Sort out control and data protection first: which data may go in, who checks the result, who decides?
- AI makes mistakes. A person reads everything that goes to customers first.
- Small companies use AI much less often than large ones, and in the trades it is still a small minority.
- Start with one task that comes up often and is easy to check.
The numbers: small businesses use AI less often
According to the Statistisches Bundesamt (Federal Statistical Office), just over a quarter of companies in Germany used AI in 2025 (26%). The gap by size is clear:
| Company size | Share using AI (2025) |
|---|---|
| 10 to 49 employees | 23% |
| 50 to 249 employees | 36% |
| 250 or more employees | 57% |
The digital association Bitkom also sees a clear trend: among companies with 20 or more employees, AI use rose from 20% to 36% within one year.
The trades look different. In the Bitkom study "Digitalisierung des Handwerks" (Digitalisation of the trades; 504 businesses surveyed, 2025), 4% said they already use AI. 84% say AI is not a topic in their business.
What this means for you: you are not too late if you don't use AI yet. But it is worth checking calmly where it could take work off your hands and where it can't.
First: control and data protection
In the same Bitkom study, 96% of trade businesses name concerns about IT security and data protection as an obstacle to digitalisation. That concern is justified. So these questions belong at the start, not at the end:
- Which data may go in? Customer data does not belong in free AI tools. Use services with a data processing agreement and, where possible, servers in the EU.
- Who checks the result? AI sounds convincing but is sometimes wrong. Quotes, prices and commitments are always checked by a person before they go out.
- Who decides? AI makes suggestions. Whether you accept a job, name a price or respond to a complaint is up to you.
- What does the law say? The EU AI Act requires, among other things, AI literacy in your team. What exactly applies is explained in our article on the AI Act.
Then and now
| Task | Before | Today with AI |
|---|---|---|
| Customer enquiries | Call back when there's time | Enquiries are pre-sorted and summarised, you decide who gets an answer first |
| Quotes and emails | Written by hand every time | Draft from bullet points, you check and send |
| Translations | Translation agency or not at all | Draft in many languages, have important texts checked |
| Receipts and invoices | Typing them up | Read automatically and prepared for checking |
| Analysis | Excel skills needed | Ask a question in everyday language |
| Social media | Agency or your evenings | Ideas and draft texts that you adapt |
8 areas of use for small businesses
1. Pre-sorting customer enquiries
An AI assistant on your website answers common questions about services, service area and process, and takes enquiries with the details you need, even after hours. A clear handover matters: whatever the assistant doesn't know for sure, it passes on to you instead of guessing.
2. Quotes, emails and texts
Bullet points in, draft out. Whether it's a quote text, a payment reminder or a reply to a complaint: AI provides a suggestion. You check it, adjust it and send it. Especially with quotes that get written at the kitchen table in the evening, this can take typing off your hands. Prices and quantities remain your call.
3. Marketing and social media
Ideas for posts, texts for before-and-after photos from the job site, drafts for replies to reviews. It doesn't replace real photos of your work, but it helps with the wording. Deepfakes, meaning AI content that resembles real people, places or events and falsely appears genuine, must be labelled under the AI Act.
4. Accounting and documents
AI reads invoices, delivery notes and receipts, recognises amounts and dates and prepares them for bookkeeping. That saves typing. You should still check the results, especially at the beginning.
5. Phone
Voice assistants can take calls, answer standard questions and note callback requests, and some also enter appointments into a calendar. But there are specific rules on the phone, such as telling callers it's an AI and the issue of recordings. What you should clarify first and which simpler solutions exist is covered in our article on AI phone assistants.
6. Knowledge in the business
An internal assistant that knows your price lists, material lists and processes. New staff ask: "How do we bill an emergency call-out?" and get the answer from your own documents. This only works as well as those documents are up to date.
7. Multilingual communication
Customer emails, quote texts or information on your website in other languages, without a translation agency. For contracts and important commitments, it's worth having someone who speaks the language check it.
8. Analysis
"Which service did we bill most often last year?" or "In which month do we get the most enquiries?": with AI you can query your spreadsheets in everyday language. You should spot-check the figures.
Examples from different industries
- Trades: quote texts from measurement notes, pre-sorting enquiries with photos, photo documentation with a draft report.
- Cleaning, gardening and other service businesses: capturing enquiries with property size and preferred date, summarising job information for the team, reading invoices.
- Office and services: minutes, email drafts, reading and filing documents.
- Hospitality, practices, retail: replies to reviews, texts in several languages, pre-sorting enquiries (in medical practices with particularly careful data protection). For appointments and reservations there is often ready-made industry software that is enough.
Why small businesses can start quickly
Large corporations need coordination and projects for every rollout. A small business can start with a single task and decide after a few weeks whether it helps. If you can't find an extra office worker, AI can support routine tasks. But the responsibility stays with you.
What you can skip
- Tasks that aren't worth it. Something that takes five minutes once a month doesn't need automating.
- AI for problems that don't need AI. If enquiries arrive incomplete, a form that asks for the necessary details often already helps.
- Big promises. Anyone who guarantees you fixed time savings or more revenue through AI can't know that in advance.
How to get started in 4 steps
- Note your time drains: for one week, write down which tasks keep repeating and which office work is left until the evening.
- Pick one task: ideally one that comes up often and whose result is easy to check, such as draft replies to standard enquiries.
- Set rules and bring your team along: which data may go in, who checks? Add a short training session, and take questions seriously.
- Assess honestly after a few weeks: if it helps, the next task is up. If not, drop it.
Tip: In our practical examples on the blog you can see, using made-up businesses, what processes like these can look like.
How we can help
We help founders, self-employed people and businesses present themselves clearly online and handle customer enquiries in an orderly way. We use AI where it helps. In the area of automation we connect enquiry, draft reply and filing so that you stay in control. And if a simple solution without AI is enough, we'll tell you.
In a free initial consultation, we look together at where time is lost in your business today and whether AI is the right answer at all.
Sources & further reading
- Statistisches Bundesamt: Unternehmen mit Nutzung von KI nach Beschäftigtengrößenklassen (2025) destatis.de
- Statistisches Bundesamt: Jedes fünfte Unternehmen nutzt künstliche Intelligenz (2024) destatis.de
- Bitkom: Durchbruch bei Künstlicher Intelligenz (September 2025) bitkom.org
- Bitkom: Digitalisierung des Handwerks, Studie 2025 (PDF) bitkom.org


