How much does an AI project cost for an SME in Switzerland?

It’s the question that comes up most often — and the honest answer is: “it depends on the scope”. An AI project has no single price, because it covers very different realities, from a simple assistant to a platform connected to several pieces of software. Rather than an abstract “AI budget”, it’s better to reason use case by use case, as we do in our AI solutions.

Without promising a magic figure, we can clarify what makes the cost vary and give useful orders of magnitude to frame a budget.

What makes the price vary

Four factors explain most of the gap between a small and a large project:

• Scope — an assistant that answers frequent questions costs far less than a bespoke business tool.
• Integrations — connecting to a CRM, to invoicing or to a calendar requires specific development.
• Data preparation — structuring and cleaning your documents for a knowledge base represents a real part of the work.
• Support — training teams, adjustments and ongoing follow-up over time.

Some orders of magnitude

To set expectations without inventing a price: a simple chatbot, connected to your content, starts at around a few thousand francs — we detail this specific case in our article on the price of an AI chatbot for business. An automation project or a bespoke tool is priced according to its scope and its integrations. The logic is the same as for a website: the price follows the scale of the need, not a fixed grid.

Initial investment and recurring costs

An AI budget has two parts not to be confused: the build (design, development, integration) and the running costs (hosting, model usage, maintenance, evolutions). Recurring costs generally remain modest relative to the time freed up, but they exist — better to anticipate them from the outset to avoid nasty surprises.

Build bespoke or subscribe to a tool?

The choice between bespoke development and a subscription to an existing solution weighs heavily on the budget — in the short and the long term. A SaaS tool limits the upfront investment but involves a recurring cost and less control; bespoke costs more to build but pays off if it becomes central. We compare the two in our article bespoke AI tool or SaaS solution.

How to keep the budget under control

The best way to control costs is to proceed in stages: start with a high-impact, low-risk case, measure the gain, then reinvest. That way you avoid the costly, uncertain “big project” in favour of a gradual, steered investment. A preliminary audit lets you price your specific case precisely rather than reasoning on averages.

To judge an AI budget, then, don’t look only at the quote, but at the ratio between the cost and the time — or the customers — the project saves you. A modest outlay that frees up several hours each week pays off quickly; an impressive but unused tool is expensive whatever its price. The right reflex is to tie each budget line to an expected, observable benefit, then to verify that benefit once the project is live.