AI & automation, built around your data
We help companies use AI and automation where it actually pays off. Not a strategy deck — working tools your team uses on Monday morning. In our experience the value hides in two places: processes that run on habit, and data that nobody can see. We start there.
There are four ways to work with us. They are not steps in a programme. Start with whichever one fits, combine them, or stop after any of them.
AI & automation training
Hands-on sessions built around the work your team already does: quotes, invoices, reporting, customer replies, hiring admin. People leave knowing which tools help, which do not, and where the quick wins are in their own week. This is often how a first conversation starts.
There is a second reason to train, and it is less comfortable. Your team is almost certainly using AI already — pasting a client email or a contract into a free chatbot to get through the day faster. Nobody means harm, but that is company data leaving the business without anyone deciding it should, or even knowing it happened. The industry calls it shadow AI. Training replaces that hidden usage with tools you chose and ground rules everyone understands.
- Your team is curious about AI but nobody has time to sort hype from help
- You suspect people already paste work into chatbots and nobody has set rules
- You want quick wins before committing to a project
- Tools your team starts using the same week
- Ground rules for what may go into which tool, and what must not
- A shared, realistic picture of what AI can and cannot do for you
AI audit
A close look at how your business actually runs. We interview a few people, watch the work happen, and write down where time leaks away. Just as important, we map the data your business produces — the numbers and histories trapped in inboxes, spreadsheets, PDFs and people’s heads. That data is usually worth more than any tool we could add on top.
You get back a short, ranked list of what is worth automating, with a rough cost and saving next to each item. No obligation to build with us. If nothing is worth doing, we tell you.
- You feel the time leaking but cannot point at exactly where
- You want numbers before you commit to anything
- A ranked list of automation candidates, costed
- A map of the data your business produces, and what it could be worth
Implementation
We build the automations for you. Rules and plain plumbing first — tools like n8n that do exactly the same thing every single time, because that consistency is what a business runs on. AI is different: ask a model the same question twice and you can get two different answers. So we use it only where that trade is worth it, such as reading messy documents or drafting first replies, always with a rule or a human checking the output. It is the belief that runs through everything we do: as little AI as possible.
Wherever we can, it runs on your own infrastructure. And we start with internal back-office work, where a mistake is cheap and easy to catch, before anything that touches your customers.
- A process eats hours every week and everyone knows it
- You already know what to automate and need it built properly
- An automation running on real work, measured against the manual way
- Documentation your team can actually read
Ongoing advisory
Once the first automation is earning its keep, we go further: the next process on the list, the harder edge cases we set aside, and the new possibilities that open up as your data becomes usable. We measure against the numbers we agreed at the start, so you can see what you are getting.
- You want a partner across quarters, not a one-off project
- The first wins landed and the list of ideas keeps growing
- A running automation programme, reviewed against agreed numbers
How an engagement works
Every project follows the same shape, and every step ends with something concrete you can read back to us:
- We find what should be automated by watching the work, not by starting from a wish list.
- We pick a first target that is low-risk and quick, even if it is not the biggest one, because the point is to prove the approach.
- We map the process as it is today, then as it should be, on one page you can read back to us.
- We prove the tricky part on your real data before quoting a full build, and we are honest that AI systems start around 70 to 80 percent and improve from there.
- We launch on a small slice first, with a human checking the output, and measure the result against the manual way.
No step commits you to the next one.
We are based in Nyon and work across Suisse romande — Vaud, Geneva, Lausanne, Morges, Rolle and the wider Lake Geneva region — where meeting in person is easy. For everyone else, we work remotely. The method travels well.
Common questions
Is my company too small for this?
Probably not, and no company is too big either. The best first automations are small: one repetitive task that quietly eats someone's afternoon. The method is the same whether you have four people or four hundred. If there is nothing worth doing, we tell you.
What does it cost to start?
A first conversation is free. If it makes sense to go further, we scope the work together after a short call. We do not publish packages, because every business runs differently and the price should follow the work, not a menu.
Where does my data go?
Ideally, nowhere. We build automations to run on your own servers whenever we can, so your data never leaves your company. If parts of your stack already run elsewhere, we recommend Swiss or EU alternatives. emod can host things for you when that is easier, but the principle stays the same: your data belongs inside your company.
Why not just use AI for everything?
Because AI is not deterministic. Ask a model the same question twice and you can get two different answers, which is fine for a draft email and risky for a price, a payroll run or a stock count. Rules-based automation does the same thing every time and can be audited, so that is what we use where consistency matters. AI earns its place on the fuzzy parts, like messy documents and first drafts, with a rule or a human checking the output.
How long before we see a result?
The audit takes a couple of weeks. A first automation is deliberately small and low-risk, so it usually goes live in weeks rather than months, measured against the manual way of doing the same work.
Do we get locked in?
No. We build with open, widely used tools rather than anything proprietary to us, and wherever possible it all runs on your own infrastructure. Another provider could take over the work at any point.
Do we need to understand AI already?
No. Part of the job is making it understandable. We use rules where work has to happen the same way every time, AI where judgement or messy input is involved, and we explain the difference in plain language as we go.
Which areas do you serve?
We are based in Nyon and work across Suisse romande — Vaud, Geneva, Lausanne, Morges, Rolle and the wider Lake Geneva region — where meeting in person is easy. Beyond that we work remotely, with clients anywhere in the world.