When a small business starts using AI in customer-facing work, a quiet question follows: do we need to tell people? For many owners, the instinct is to say nothing, because the AI is "just a tool" and a person is still responsible. That instinct is understandable, but it misreads both the rules and the mood of customers. In Switzerland, being open about AI is increasingly expected, and done well it can strengthen trust rather than weaken it.

What customers say they want

Swiss attitudes are fairly clear. The AXA Cybersecurity Monitor 2026, a survey of 1,490 people carried out with the Sotomo research institute and reported by Swissinfo, found that 94 per cent favour mandatory labelling of AI-generated content. The same survey found that 87 per cent expect cyber fraud cases to increase. People are uneasy about synthetic content, and they want to know when they are dealing with it.

For a small firm whose business rests on personal relationships, that is a signal worth taking seriously. Customers who discover undisclosed AI use later may feel misled, even if nothing went wrong.

What Swiss law already expects

Switzerland has no dedicated AI act, but that does not mean there are no rules. The FDPIC has stated that the Data Protection Act applies directly to AI-supported processing. In its guidance, it draws out several points that bear directly on customer communication:

  • The purpose, functionality and data sources of AI-based processing should be made transparent.
  • People have a right to know whether they are communicating with a machine, and whether the data they enter is used to improve self-learning systems.
  • Tools that manipulate the faces, images or voices of identifiable people should be clearly indicated.
  • People can object to automated processing and ask for a human to review automated individual decisions.

More may follow. In February 2025, the Federal Council decided to bring the Council of Europe's AI convention into Swiss law, with changes made sector by sector where possible. Swissinfo reported that transparency is one of four areas where legal measures are to be defined, and that a consultation draft is due by the end of 2026.

The EU dimension

Many Swiss SMEs sell to customers in the EU, so the EU AI Act matters too. The European Commission states that people should know when they are dealing with a machine, that certain AI-generated content such as deepfakes must be clearly labelled, and that these transparency rules took effect in August 2026. A July 2026 analysis by the law firm Stibbe dates the transparency obligations to 2 August 2026, and notes that some duties fall on the provider of the AI system while others, such as deepfake disclosure, fall on the business using it.

Whether and how these obligations reach a particular Swiss company depends on what it does and where its customers are. If you serve EU customers with chatbots or generated content, that is a question for a short conversation with a legal adviser.

A practical disclosure approach

Most SMEs do not need elaborate notices. They need a few consistent habits:

  1. Say it at the point of contact. If customers chat with an assistant, say so in its first message and offer a clear way to reach a person.
  2. Update your privacy notice. Describe in plain language which tasks use AI, what data is involved, which provider processes it and where, and whether inputs are used for training.
  3. Label synthetic media. If you use generated images, video or voice in marketing, mark it. Never create a synthetic voice or likeness of a real employee or customer without consent and a clear label.
  4. Keep a human route for decisions. Where AI influences an outcome that matters to a customer, such as a quote, eligibility or prioritisation, make sure a person can review it on request.
  5. Be proportionate about drafting help. An email drafted with AI support and reviewed by a named employee is different from an automated reply. A general statement in your privacy notice may be enough for the first; the second deserves a visible note.
  6. Write it down internally. A short register of where AI is used makes it easy to answer a customer who asks.

Transparency as a selling point

There is a commercial argument here as well. Customers are hearing more about AI-driven fraud and deepfakes. A business that can say plainly which tasks AI helps with, that a person checks what goes out, and that customer data stays with vetted providers gives those customers a reason to trust it.

The goal is not a disclaimer on every page. It is a simple, honest account of how you work, available when someone wants it, and consistent with what you actually do.

Not sure which of your AI uses need a disclosure? The free AI Readiness Audit includes a review of customer-facing touchpoints and data flows.

Sources

  • Swissinfo (Keystone-SDA), "Swiss citizens concerned about AI cyber fraud", 20 April 2026. link
  • Federal Data Protection and Information Commissioner (FDPIC), "Update - Current data protection legislation is directly applicable to AI", 8 May 2025. link
  • Federal Data Protection and Information Commissioner (FDPIC), "AI and data protection", 24 September 2025. link
  • Swissinfo (Keystone-SDA), "Swiss government wants to adopt Council of Europe AI conventions", 12 February 2025. link
  • European Commission, "AI Act", last updated 19 October 2026. link
  • Stibbe, "The AI Act's Transparency Obligations: Rules, Scope and Timeline", 27 July 2026. link