Before buying licences or commissioning any development, an SME needs to know where the gains really are, not just where it has been told they are. Finding out does not take a fifty-page questionnaire or a three-month engagement: half a day of work with the people who do the tasks is enough. Here is what we look at, in what order, and what you have in hand at the end.

What we look at during the session

Four topics, in this order:

  • Recurring tasks. Data entry, reminders, letters, quotes, meeting notes, answers to the same questions. For each one: how often per week, how many minutes, and who does it.
  • The data involved. What each task would put into an AI tool, sorted into three levels: internal, client and protected. Our security guide explains these levels.
  • Tools already in use. In many SMEs, staff are already using a free AI tool on a personal account. We say so openly, without looking for someone to blame, and decide what moves into a business tool and what stops.
  • The current cost. The time spent on the tasks in question, first in hours, then in francs. Without that figure, the gain from AI is only a feeling.

Participants then try AI on their own letters, quotes and files. This is also the initial training: what the tool does well, what it makes up, and how to check.

How we rank the use cases we find

Each idea is scored on three criteria, the gain / effort / risk grid we use everywhere:

  1. Gain: hours per week, measured on real work.
  2. Effort: the assistant on its own (level 1), batch processing (level 2), or a connection to your software (level 3).
  3. Risk: does a mistake stay in-house, reach a client, or involve protected data?

We keep three use cases: the best balance of gain and effort, with low risk, or medium risk if someone reviews the output. Not the most spectacular idea. Our guide on finding the three right use cases shows the grid applied to an example.

What you have in hand at the end

  • A written action plan: three priority use cases, the tools involved, the next steps and who owns them.
  • A set of ready-to-use instructions (prompts) for your tasks.
  • A one-page sheet on which data may and may not go into the tool.

The first use case in the plan becomes a 30-day pilot, described in our guide to getting started with AI in an SME.

What the session does not promise

No "digital transformation", no shopping list of tools, no eighteen-month timetable. The aim is to decide whether AI has a place in your work, and where. If the answer is "not now, not on this process", that is a good answer: it has cost you half a day rather than several months of development.

Why not just a questionnaire

Because the gains are hidden in the real work, not in what people think the work involves. A questionnaire fills in boxes; a session with the team reveals that the "already quick" data entry actually takes an hour a day, or that three people each rewrite the same report on their own.

What it costs, and what comes next

The AI Workshop costs CHF 1,900 excluding VAT for a 4-hour half day, or CHF 2,900 for a 7-hour full day, per group of up to 20 participants, preparation and materials included. The price is confirmed before the session. The workshop stands on its own: setting up a tool for the whole team falls under AI Adoption Support, and connecting AI to your software under AI Engineering, each quoted separately.