5 practical ways to use artificial intelligence in an SME

In short

• AI already saves SMEs hours on repetitive tasks: replying to customers, preparing documents, writing content, organising and analysing.
• The real gain comes from choosing one or two well-defined tasks, not from “using AI for everything”.
• Customer data calls for care: GDPR, contracts with suppliers and, since August 2026, EU transparency rules for chatbots.

Artificial intelligence is no longer just a topic for large companies. Today, a restaurant, a building firm or an accountancy practice can use affordable tools to save time every week.

The problem is knowing where to start. There are new tools every day and too many promises. In this article we stick to the concrete: five uses we see working in small businesses, what it takes to get them going and the precautions to take.

1. Replying to customers (website, WhatsApp, e-mail)

Much of the messaging an SME receives is repetitive: opening hours, prices, availability, “do you offer this service?”, “where are you based?”. Answering everything by hand takes hours, and in the evening or at the weekend, nobody replies.

What AI can do

• An assistant on your website or WhatsApp that answers frequently asked questions based on the company’s information.
• Collect the details of a request (type of service, address, dates) and pass it on to a person, already organised.
• Suggest draft replies to e-mails, which someone reviews before sending.
• Reply in the customer’s language: Portuguese, French, German or English. In Luxembourg, that alone is an advantage.

Precautions

• The assistant should only answer on the basis of information the company has validated. If it does not know, it hands over to a person.
• Tell customers clearly that they are talking to an automated assistant. As well as being good practice, it is an obligation: the transparency rules of the EU Artificial Intelligence Act (AI Act), applicable since August 2026, require people to know when they are interacting with an AI, unless this is obvious.
• Do not promise what only a person can guarantee, such as final prices or project deadlines.

2. Documents and quotes

Quotes, proposals, standard contracts, site visit reports: documents that almost always follow the same structure, but that somebody has to write.

What AI can do

• Turn quick notes (or a voice message after a site visit) into a structured quote, based on your templates and price list.
• Summarise long documents: a specification, a supplier contract, a letter from the authorities.
• Translate and adapt documents into another language, with a human review at the end.

Precautions

• Prices and terms always come from your price list, never “made up” by the AI.
• Review before sending is a must. AI can be confidently wrong: a transposed figure in a quote is expensive.
• Legal and tax documents still need to be validated by a professional.

3. Marketing content

Social media posts, website copy, newsletters, product descriptions: tasks that many business owners put off for lack of time.

What AI can do

• Suggest a posting calendar based on your services and the time of year.
• Write first drafts, which you then adjust to your own tone.
• Adapt the same content for several channels and several languages.
• Generate ideas for headlines, FAQs and answers for your website.

Precautions

• Generic content does not sell. AI writes the draft; your experience, your real case studies and your photos make the difference.
• Check all facts, figures and legal references.
• Be careful with generated images: they must not mislead customers about products or work you have not done.

4. Internal organisation

This is where many SMEs save the most time, and where it is least talked about.

What AI can do

• Summarise meetings and turn the discussion into a task list with owners.
• Sort incoming e-mails and suggest priorities.
• Extract data from supplier invoices and organise it for the bookkeeping.
• Answer team questions based on internal procedures (“how do we place an order with supplier X?”).

Precautions

• Start with one task and measure the time saved before expanding.
• Decide who reviews what: automating without supervision in a process involving money is risky.
• Make sure the team knows how to use the tools. Since February 2025, the AI Act has required companies that use AI systems to take measures to ensure a sufficient level of AI literacy among their staff.

5. Analysing figures and customers

Many SMEs have data (sales, enquiries, reviews, response times) that nobody looks at, because there is no time or knack for spreadsheets.

What AI can do

• Read a sales export and answer simple questions: which service is growing, which month is weakest, which customers have stopped buying.
• Analyse dozens of reviews and summarise what customers praise and criticise.
• Compare quotes sent with quotes accepted and suggest where business is being lost.

Precautions

• The conclusions are only as good as the data. Check the most important figures.
• Use AI to ask better questions, not to make decisions on its own.

Risks and GDPR: the essentials

Using AI with customer data means processing personal data. The GDPR rules apply just as they do to any other tool.

Before you start:

1. Know where the data is stored. Choose providers with servers in the European Union, or with appropriate safeguards for transfers outside it.
2. Read the terms. Check whether the provider uses your data to train its models and whether you can switch this off. Business versions usually offer more safeguards than free ones.
3. Sign a data processing agreement (DPA) with the provider when it processes personal data on your behalf.
4. Minimise. Do not paste more data into the tool than you need. Identification numbers, health data or bank details should not go into a general-purpose tool.
5. Inform. Update your privacy policy and tell customers when they are talking to an automated assistant.
6. Keep a human in the loop for decisions that affect customers: prices, refusals, complaints.

The data protection authority in Luxembourg is the CNPD, which publishes useful guidance for businesses.

Where to start

The worst way to start is to buy a tool and then look for a problem for it. The best is the reverse:

1. List the week’s repetitive tasks and how long they take.
2. Choose one, with a clear impact and low risk.
3. Test it for a month and measure the time saved.
4. Only then extend it to other tasks.

A well-scoped AI project may be eligible for public support, such as SME Packages AI. See our SME Packages Digital guide to understand how these packages work.

Next step

The free AI Opportunities framework helps you map your company’s tasks and identify where AI delivers the most benefit with the least risk. You can also order it already filled in by an In Digi consultant (AI-assisted) for €49.

If you would rather draw up this map with us, book a one-hour consultation (€100 + VAT, deducted if you hire us within 90 days). We look at your processes, suggest a first use case and explain what it takes to get it running safely.

Sources

• European Commission – Regulatory framework for AI (AI Act): application timeline
• Regulation (EU) 2024/1689 (AI Act)
• CNPD – Commission nationale pour la protection des données
• Guichet.lu – SME Packages

General information only; it does not replace professional advice. Always check your situation with an accountant or lawyer.