AI for the self-employed and small businessesFive important tasks you can hand over

The day is full before it has even begun. Customers want a quote, enquiries are waiting for a reply, nothing has gone up on Instagram for weeks, and the receipts for the bookkeeping are sitting in a shoebox. If you have a team, there is the added problem that everyone writes differently and the know-how rests with two or three people. Everyone is talking about artificial intelligence, but nobody tells you what to start with on Monday.

Why now is the right time

Austrian businesses have picked up speed on AI. In 2025, 30 per cent of companies with ten or more employees used AI technologies, compared with 20 per cent a year earlier; the EU average was 20 per cent (Statistik Austria, press release of 24 June 2026). Among smaller companies with ten to 49 employees, which includes most businesses with up to 30 staff, the share was 26.2 per cent, and AI is most often used in marketing and sales, by 45 per cent of users (Statistik Austria, IKT-Einsatz in Unternehmen 2025).

There is no official figure for one-person businesses, because the survey only starts at ten employees. Yet they are the majority: the Austrian Economic Chamber (Wirtschaftskammer, WKO) counts 376,112 one-person businesses, which is 62.1 per cent of its members (WKO, EPU-Factsheet, as at December 2025). A survey for the WKO shows what holds many people back: among users, 39 per cent name the legal framework as an obstacle and 37 per cent concerns about confidential data (MARKET for WKO, KI und Robotik aus Sicht der Unternehmen, survey December 2025 to January 2026).

I have been in sales for almost forty years and have brought many new products and technologies to market. One thing has always held true: new technology catches on when it makes a job easier that people do every day anyway. That is why I never start with the tool, but with the question of what takes up most of your time.

Five tasks to start with

Quotes and replies to enquiries. Take a master painter who answers a dozen enquiries every week. She gives the AI a few keywords on what the customer wants, which surface area, which time frame, which special requirements, and gets back a friendly, complete draft in her own tone. She fills in the price herself. That way the draft is ready in a few minutes, and the reply can go out the same day. In a business with an office, everyone works from the same template, so that every customer gets the same quality, no matter who replies.

Social media posts. A photo from the building site, three keywords to go with it, and the AI turns them into a short post for Instagram or Facebook. If you like, you can have it suggest a monthly plan with four topics at the same time. What matters is that the post sounds like you. To make sure it does, you give the AI three examples of your own writing once.

Replies to reviews. Good reviews deserve a thank you, bad ones a calm, factual reply. With an angry review in particular, a draft that was not written in the heat of the moment helps. The AI makes a suggestion, you decide what gets published. How reviews and online visibility are connected is at the heart of VISO SCALE in our SCALE projects.

Preparing the bookkeeping. AI does not replace your tax adviser. But it can sort receipts by month and type, word the text on invoices, prepare questions for the accountancy firm or draft friendly payment reminders from a list of unpaid invoices. For receipts containing customer data, the rule I describe further down applies.

Preparing for and following up on appointments. Before a meeting with a customer, you have the AI put together the most important questions. Afterwards you dictate your notes and get a summary with the next steps and a draft of the email to the customer. In my experience, this is the task that most often gets left undone, and it is precisely this one that often decides whether a conversation turns into an order. In a team, the same applies to meetings and handovers: a short summary means nobody has to ask who is taking care of what.

If you work in sales and need content in your own voice, we offer VOCO SCALE Independent, an AI assistant that prepares texts in your tone and sends nothing without your approval.

What changes when you have a team

As soon as several people in the business work with AI, it is not enough for the boss to know what is allowed. In Germany, in businesses with fewer than 50 employees, 33.5 per cent of AI users use tools that the employer has not provided, mostly private accounts (Institut der deutschen Wirtschaft, IW-Kurzbericht Nr. 63/2026, 19 September 2026). The risk: customer data ends up in other people's systems without any control.

The solution is simple and takes little time: one approved tool for everyone, one page of rules on what may go in and what may not, and a short training session for everyone who works with it. From around twenty staff, it is also worth securing the business's know-how in an orderly way, so that it does not retire along with one person; for this, the SCALE projects offer RENVO SCALE.

Data protection in plain words

As soon as names, phone numbers, addresses or invoices of your customers are involved, the General Data Protection Regulation (GDPR) applies. The Austrian Data Protection Authority (Datenschutzbehörde) makes clear that the company using an AI system is, as a rule, itself the controller under data protection law, even if the technology comes from a provider (Datenschutzbehörde, Künstliche Intelligenz und Datenschutz, accessed 23 October 2026).

This leads to three simple rules. First: free accounts are fine for trying things out, but not for customer data; the WKO expressly advises against entering sensitive data there (WKO, Datenschutzkonforme KI im Betrieb, 15 May 2025). Second: for working with customer data, you need a business plan for which the provider concludes a data processing agreement (Auftragsverarbeitungsvertrag) under Art. 28 GDPR, ideally with data processing in Europe; many individual plans for private users do not offer this. Third: switch off the setting that allows your inputs to be used to train the model. If in doubt, replace names in the text with placeholders such as Customer A.

What the AI Act requires of you

The European AI Act also affects small businesses. Under Article 3 point 4, a deployer is anyone who uses an AI system under their own authority; the only exception is personal, non-professional use. So anyone who uses AI for their business is a deployer.

Article 4 on AI literacy has applied since 2 February 2025. It was amended with effect from 27 July 2026 by Regulation (EU) 2026/1744: deployers must now take measures to support the AI literacy of their staff, but they do not have to guarantee a particular level of knowledge for each individual (EUR-Lex, Regulation (EU) 2026/1744). The AI Service Centre (KI-Servicestelle) of the Austrian regulator RTR (KI-Servicestelle) stresses that the appropriate measures depend on the AI system in use and recommends documenting them (RTR, KI-Servicestelle, KI-Kompetenz). For a small business, this means: a short training session for yourself and the team, a few written rules and a note of who was trained and when.

And one more point, in case you are planning a chat on your website: since 2 August 2026, people must be able to recognise that they are talking to an AI system (Article 50 AI Act).

Where the limits are

AI writes quickly and convincingly, but not always correctly. Figures, prices, deadlines and legal questions are things you check yourself or have checked by an expert. The AI does not know your customers the way you do, and it bears no responsibility. That is why approval stays with you, every time.

A plumbing and heating business with twelve staff

The following case is an illustrative example from practice.

A plumbing and heating installer in Vorarlberg runs his business with twelve staff: nine on the building sites, two in the office, plus the boss. Enquiries come in by phone, by email and through Google, the boss writes the quotes in the evening, and reviews mostly go unanswered. In the office, one colleague already uses a private AI account, the other none at all. Every week, many hours are lost to follow-up questions and reminders.

The turning point is the decision to start with a single task: replying to enquiries. The office gets a business plan with a data processing agreement, a template with three sample texts in the business's tone and one hour of training. There is one page of rules for all twelve, and the private account is no longer used for customer data. Only once that is running smoothly are reviews and social media posts added.

The goal is for every enquiry to be answered the same day, for the boss to have his evenings back and for the business to stay visible online, without anyone having to put in extra hours for it.

What you can do from tomorrow

• For one week, write down what you need desk time for, and choose the task that takes up the most time.
• Try out this task with a free account, using only made-up sample data and three of your own texts as a model for your tone.
• Ask your team who already works with AI and with which account.
• Find out which business plan with a data processing agreement is suitable for your real customer data.
• Write down in two or three sentences which rules apply to AI in your business and who has been trained.

If you run a hotel or a guesthouse, also read my article AI in the family business.

Frequently asked questions

As a one-person business, do I even need AI literacy under the AI Act?

What is settled: anyone who uses AI for work is a deployer under Article 3 point 4 of the AI Act. What is still open is how far Article 4 reaches when nobody is employed, because the provision refers to staff and to persons acting on behalf of the deployer. A short, documented training session is therefore a sensible precaution, at the latest as soon as you employ or engage someone.

Does this also apply to my business with employees?

Yes, and here the position is clearer than for a one-person business: Article 4 of the AI Act refers expressly to staff. A short training session per person, one page of rules and a note of who was trained and when form an appropriate framework for a business with up to 30 staff.

May I enter customer data into ChatGPT or similar programmes?

Not into free private accounts. For customer data, you need a business plan with a data processing agreement under Art. 28 GDPR and training switched off. If in doubt, replace names with placeholders.

Is AI too expensive for a small business?

The current prices are set by each provider. What matters is whether the time you save is greater than the effort for setting up and training. That is why you start with the task that takes up the most time.

I am not a technical person. Can I manage this?

Yes. You operate most tools with perfectly ordinary sentences. What helps at the start is a short training session and someone you can reach in the first few weeks when a question comes up.

Will AI replace my tax adviser or my bookkeeper?

No. The AI can prepare, sort and word things. Tax assessments and annual accounts remain with your accountancy firm.

If you would like to know which task would make the biggest difference in your business, let us talk about it. The first conversation is free of charge and without obligation. You can see how things continue afterwards in seven steps under How I work.

This article is general information and does not replace legal or tax advice in an individual case.