Customer service is often the first place a business considers AI, and for good reason: the volume is high, many questions repeat, and customers expect answers outside office hours. It is also the place where a poorly judged deployment does the most visible damage. The evidence from the past eighteen months points to a fairly clear line between what customers accept and what they resent.

Customers accept speed on simple questions

Swiss consumers are not hostile to chatbots. A Comparis survey reported by Swissinfo in 2024 found that 72% of respondents would prefer a chatbot for questions about delivery status, and the same share favoured chatbots for faster searches online. A later Comparis survey, covered by Swissinfo in April 2026, found that 76% of adults in Switzerland now use AI tools in daily life and that 61.9% considered written correspondence with a chatbot an interesting option.

The same survey also marked the limits. Respondents were wary of sharing personal data with AI, especially health information, and 55.5% said they would never confide psychological problems to a chatbot. Acceptance is high for transactional questions and falls sharply as questions become personal, sensitive or consequential.

Customers resent being kept from a person

The clearest international data comes from Gartner. In a June 2025 survey reported by The Register, 51% of customers said they trusted human agents most to resolve a service issue, against 7% who trusted AI most. Being able to reach a person ranked as customers' second-highest priority in a service interaction, and 62% worried that AI would make that harder.

Businesses have adjusted. In the same research, half of the 163 service leaders surveyed said they would abandon plans to significantly reduce their customer service staff by 2027, and 95% planned to keep human agents alongside AI. Gartner's analysts also noted that firms often underestimated the total cost of running generative AI, which made the expected savings hard to realise.

Swiss SMEs appear to share the caution. An AXA study reported by Swissinfo in September 2026 found that while AI use for process optimisation and data analysis has grown, customer-facing uses such as customer service have stagnated, with the authors pointing to limited trust in AI for direct customer contact.

Your chatbot speaks for your business

The case most often cited on liability is still Air Canada's. In 2024, a Canadian tribunal ordered the airline to compensate a passenger after its website chatbot wrongly told him he could claim a bereavement fare refund after travelling. As CBS News reported, the airline argued the chatbot was responsible for its own statements. The tribunal disagreed, finding that the chatbot was "still just a part of Air Canada's website" and that customers should not have to work out which parts of a company's site are accurate.

The principle travels well. If your assistant quotes a price, a delivery date or a warranty term, a customer will reasonably treat it as your word.

Transparency is now also a legal matter in the EU. Since 2 August 2026, the AI Act has required that users be clearly told when they are interacting with an AI system rather than a person, with chatbots and AI agents named as examples. Swiss firms whose chat tools serve customers in the EU should check whether the obligation applies to them. Disclosing it is good practice either way.

A design checklist for SME customer service

  • Start with the questions that have one right answer. Order status, opening hours, document requests, appointment changes. These are where customers already prefer speed.
  • Ground answers in approved sources. The assistant should answer from your current price lists, terms and FAQs, not from general knowledge. When the source does not cover the question, it should say so.
  • Make the route to a person obvious. One click or one phrase, offered early, not after three failed attempts. Hand over the conversation history so the customer does not repeat themselves.
  • Keep sensitive topics human. Complaints, health, finances, cancellations and anything involving a dispute go to a person by default.
  • Never let the assistant make commitments it cannot keep. Refunds, discounts and exceptions to policy should be drafted for a person to approve, not promised automatically.
  • Say that it is AI. A short, plain statement at the start of the conversation.
  • Review a sample every week. Read twenty real conversations. Look for wrong answers, frustrated customers and questions the assistant should have handed over.

The practical balance

For a small Swiss business, the goal is not to replace the person who answers the phone. It is to take the repetitive questions off their desk so they have time for the conversations that build loyalty. Customers will reward a fast, honest assistant that knows its limits. They will not forgive one that blocks them from a person, or promises something the business then refuses to honour.

Want to know which of your customer enquiries are safe to automate? The free AI Readiness Audit maps your recurring tasks and ranks what is worth automating first.

Sources

  • Swissinfo, "Chatbots on the rise in Switzerland", 17 April 2024. link
  • Swissinfo, "Three-quarters of Swiss use AI in everyday life: survey", 2 April 2026. link
  • The Register, "Half of businesses rethink ditching humans for customer service bots", 11 June 2025. link
  • Swissinfo, "AI not a job killer, say most Swiss businesses", 15 September 2026. link
  • CBS News, "Air Canada chatbot costs airline discount it wrongly offered customer", 19 February 2024. link
  • European Commission, "Safer and more transparent AI", 2 August 2026. link