AI Chatbot Cost in Switzerland: What You Pay For (and What You Don't Need)

The AI chatbot cost for a Swiss business can range from a few hundred francs for a simple widget to tens of thousands for a custom assistant connected to every system in the company. Most small and medium businesses need something in between: a chatbot that answers customers correctly, captures leads and is live quickly. This article explains what drives the price, what you actually pay for, and where you can save.

What drives the AI chatbot cost

Five decisions explain most of the difference between two quotes:

1. Knowledge: a bot that answers from your website and a few documents is quick to build. One that must read live data (stock, bookings, customer accounts) needs integrations.
2. Actions: answering questions is the base. Capturing leads, booking appointments or creating tickets adds steps that must be tested.
3. Languages: in Switzerland, two or three languages are common. Each language needs its own testing.
4. Channels: website only, or also WhatsApp, Instagram or the phone (that is a voice agent).
5. Quality assurance: guardrails so the bot never invents prices or promises, plus tuning after launch with real questions.

What a CHF 990 chatbot launch includes

As a concrete reference, here is our own entry package, the AI Chatbot Launch (CHF 990, about 2 weeks):

• trained on your website and documents;
• leads sent to your inbox (name, email, what they need);
• up to 2 languages;
• one month of tuning included after launch.

You can see what such a bot feels like on your own website in our free chatbot demo: enter your address and ask it questions your customers would ask.

Running costs: what to expect every month

After the launch there are usually two small running costs:

• Hosting of the chatbot and its knowledge base.
• AI usage: the language model charges per message or per volume of text. For a typical small business website the volume is modest.

Ask any provider for an estimate based on your expected number of conversations, and who pays the AI bill. A good set-up lets you see the usage and change the model later without rebuilding the bot.

Where you can save, and where you shouldn't

Save on: extra channels and integrations at the start. Launch on your website, learn which questions people really ask, then connect more systems.

Don't save on: accuracy and testing. A bot that invents a price or a delivery date costs more in lost trust than it saves in set-up.

Build it yourself, use a template tool, or hire an agency?

There are three common routes, and each fits a different situation:

• Do it yourself with a chatbot builder: the lowest price, often a monthly subscription. Good when your questions are simple and you have time to write and maintain the answers yourself.
• A template from a platform: quick to start, but usually limited to that platform's channels, design and AI model.
• An agency set-up: you pay once for the build and get a bot trained on your content, tested for your questions, with lead capture and languages done for you. Good when the chatbot must represent your brand correctly from day one.

Whatever you choose, keep the right to export your content and conversations, so you are never locked in.

How long it takes

A focused website chatbot is usually live in about two weeks: a kick-off call to list the questions and goals, a few days to prepare the knowledge and build the bot, a round of testing with your team, then go-live and tuning. Integrations with booking, CRM or shop systems add time, which is why we often start without them and add them once the bot has proved itself.

Questions to ask before you buy

• Will the bot answer only from our content, and show where the answer came from?
• How are leads delivered (email, CRM) and who owns the conversation data?
• Which languages are included, and how is each one tested?
• What happens when the AI provider is busy or changes its prices?
• Can we change texts and answers ourselves later?

Published 5 November 2026