AI readiness for Swiss SMEs in 2026
For Swiss SMEs, an AI workflow needs to fit the nDSG, customer requirements, multilingual operations and the capacity of a small team. This article sets out those four constraints for owners and managing directors, and a way to test one workflow against them this quarter.
Take a Swiss SME with thirty employees, confidential client data and quotes going out in German and French. Its AI plan has to account for those conditions and for the people available to run the workflow.
Assess which workflows can use AI today, under Swiss rules, with the people you already have. Block Consult's AI readiness and workflow audit looks at six areas for each candidate workflow: data, system connections, process consistency, people, governance and economics. The same six areas work as a checklist for an initial assessment of your own.
The Swiss context in four constraints
- Data protection (nDSG): the revised Swiss data protection act applies fully to AI tools. Personal data in a prompt is still personal data.
- Data residency: some client contracts, sector rules and risk policies require Swiss or EU processing, and many do not. Knowing which applies to you decides your tool shortlist.
- Languages: workflows run in German, French, Italian and English, often within one inbox. Models handle this well. Your prompts, templates and review steps have to handle it too.
- Lean teams: few SMEs have a data science unit to spare. Whatever you adopt must be maintainable by the people who run the workflow.
What the nDSG means for AI tools
The nDSG does not forbid using foreign AI providers. It requires the same discipline as any other processing: know what personal data enters the tool, have a data processing agreement, and use a lawful transfer mechanism for providers outside Switzerland. That mechanism is an adequacy decision, certification under the Swiss-U.S. Data Privacy Framework or standard contractual clauses.
Most established providers offer these, so the work is to check that they are in place for the tool you plan to use. Start with a list. For the workflow you are assessing, write down which personal data would enter the tool and whose it is. Then compare the provider's agreement and transfer mechanism with that list. Have your counsel confirm the specifics for your sector.
Where data residency requirements come from
Data residency requirements come from your client contracts, your sector's rules and your own risk policy. Some of these require Swiss or EU processing and many do not, so the answer differs from company to company.
Read the contracts and sector rules that apply to the workflow, and note what each says about where data may be processed. Many Swiss SMEs run on EU infrastructure or on US infrastructure certified under the Data Privacy Framework, and some client relationships justify Swiss or self-hosted setups. Answer the question per workflow, because the answer decides which tools you can shortlist.
Plan for several languages and a small team
A shared inbox can receive German, French, Italian and English on the same day. Models handle this well, so the work lies in the material around the model. Keep an approved template in each language for anything sent to customers, write prompts that state the language of the reply, and decide who reviews output in each language. A reviewer who cannot read a reply cannot approve it.
Whatever you adopt has to be maintained by the people who run the workflow. Before choosing a tool, name the person who will update prompts and source documents and check output quality, and ask how much of their week that takes. If nobody has the time, choose a smaller workflow. Our guide to AI governance for SMEs describes how to assign owners and set review rules by risk.
Workflow candidates for Swiss SMEs
Workflow candidates to assess include extracting data from client documents in fiduciary and accounting firms, drafting quotes from technical enquiries in machine shops, turning site notes into structured reports in building services companies, and answering support questions from internal knowledge. Check the volume, the available data and the human review step before selecting one.
These candidates are document-heavy and high in volume, and in each a person can check the result before it leaves the company. Plan that check as part of the workflow from the start. Our guide to designing human review explains how.
How to test your AI readiness this quarter
- Pick one workflow. Choose the one where the most hours visibly disappear.
- Audit it. Measure volume, touch time, wait time, rework and the cost of an error. A week of interviews and one case walked end to end are enough.
- Check the constraints. Note what personal data is involved, what your contracts say about processing and which tools your team already licenses.
- Prototype with review built in. Allow two to four weeks, use representative data and make one person accountable for judging output quality.
Each step has a written result: the workflow chosen, its figures, the list of constraints and the reviewer's verdict on the prototype. Together they show whether the workflow should be built, changed or dropped.
A workflow audit identifies where time is being lost. An AI readiness audit turns those findings into a prioritised plan with budget ranges, which can then be tested in a working prototype.
Frequently asked questions
- Does the nDSG forbid US-based AI tools?
- No. It requires a lawful basis for the transfer (an adequacy decision, certification under the Swiss-U.S. Data Privacy Framework or standard contractual clauses), a data processing agreement and appropriate care with personal data in prompts.
- Do we need Swiss-hosted AI?
- Only if your contracts, sector regulation or risk policy demand it. Many Swiss SMEs run on EU infrastructure or on US infrastructure certified under the Data Privacy Framework, and some client relationships justify Swiss or self-hosted setups. The audit should answer this per workflow.
- What is a realistic first step for a Swiss SME?
- A bounded workflow audit of one process, typically one to three weeks, followed by a prototype with a built-in review step. That is small enough to finish and concrete enough to show whether the workflow is worth building.
- Can one AI workflow handle German, French and Italian?
- Models handle the languages well. The work is on your side: prompts, templates and a reviewer for each language you send to customers.
Related reading
Let's build something durable.
Tell us about your project. We reply within two business days.

