Your workflows, rethought
We do not digitise the old route. First we look at what a workflow would look like today if you could build it from scratch. Only then does the system take over the parts nobody has to do by hand.
We take over the recurring work in your business: find it, rebuild it, run it. You get the hours back without putting a single minute into AI yourself.
We organise the knowledge once and keep it current. This is where a system knows how your business really works.
AI is changing your business, nobody argues with that any more. But reading up, trying things out, tidying up and sticking with it is a full-time job, and nobody in your company has that time spare. So it gets left, and next week the day-to-day looks the same again.
Pull figures out of three systems, drop them into the same template, send. For years now, by hand.
Between sales and production, between shifts, between departments. Asking again every time, waiting every time.
When that person goes on holiday, the department stalls. That is not a staffing problem. That is a memory problem.
Paid for privately, nobody knows which data goes where. In a workshop you would not put up with that for a week.
Something was tried, then the day-to-day took over. Nobody owned it, and today nobody mentions it any more.
Working time, handovers, mistakes, waiting time. Without that figure you can neither decide nor prove that anything got better.
The end result is not a faster process. The end result is that the process is gone, and the time is yours again.
We look for the processes that repeat the same way every week. We work out with you what one of them costs over a year. Then we rebuild it so that nobody has to sit with it any more, and we keep it running when your business changes.
Coordination problems look different on a shop floor than in a fiduciary firm. Here we show where we typically start.
See all nine industries in detail
Every morning: key figures, deviations and open items from ERP and MES.
Delivery dates watched continuously, alternatives prepared before management asks.
Structured handovers, created automatically from the events of the shift.
Budget, risks and milestones of every project, up to date daily.
Based on earlier projects, in your format, prepared for review.
Before every meeting: the full history and the open items, automatically.
By branch, product group and channel, every morning.
Sell-through rates and seasonal patterns, proposals put in front of you to approve.
Lead times, quality and terms recorded systematically.
Quality evidence prepared, entirely in house.
Skills, availability, rules. Conflicts visible before you sign off.
Open source models, no data leaves the building.
Every construction project: budget, schedule and risks, kept current automatically.
Extracted, assigned and chased automatically.
Costings and past experience for new quotes.
Full context per client, the same level of knowledge for everyone.
Changes in the law watched, the impact flagged per client.
Annual accounts and standard letters prepared from the client data.
Facts, deadlines and open points out of the files.
A draft that knows your own earlier cases.
Daily, out of your systems instead of out of somebody’s head.
Drafts in the tone of the school, out of its own context.
Out of the data you already have instead of collecting it again.
Written down in a structure, not passed on in the staff room.
Out of the statutes, the regulations and earlier answers.
A draft right after the meeting.
Quarterly, out of the business in progress.
We do not digitise the old route. First we look at what a workflow would look like today if you could build it from scratch. Only then does the system take over the parts nobody has to do by hand.
ERP, CRM, accounting, mail. The system reads what is there anyway and names the source for every answer. We do not replace your tools, we make them answerable.
Every workflow gets a traffic light, sorted by how easily a mistake can be put right. Nothing that could hurt runs on its own.
Every morning: project status, open decisions and the critical figures. Out of your systems, in your format, without anyone having to ask.
The system knows your decision history and delivers papers in the format your management expects. Instead of hours of preparation.
Processes, decisions, customer knowledge, unwritten rules. Captured in a structure once, kept current, available at any time. Even when people move on or go on holiday.
The AI is the racehorse. We are the stable behind it. The horse is fast, expensive and, without care, worth nothing after six months. The work that gets it to the starting line does not happen on the track. It happens beforehand, every day, and it looks like nothing.
A horse without a stable is out in the field within six months.
Whether it is ChatGPT, Copilot or a local model: without your company context it is a brilliant new employee with amnesia and no key card. Highly intelligent and still useless. Flip the switch.
Happy to help. Here is a suggestion for your Monday briefing:
Would you like me to flesh the points out further or build a template?
Monday. Three points for management:
Correct, generic, useless. You get this from any assistant, in any company, the same every Monday. It knows neither your customers nor your figures nor your hall 2.
Your figures, with the honest part included. Other calculators only show the gross saving.
A workflow that repeats every week. A weekly report, a handover, a quote.
Above is the ideal case. Two things pull it down.
The one-off setup, the licences, and the time of your own people.
Every figure here is your own assumption, not a commitment. We say what is realistic for you once we have seen the workflow, not before.
In many Swiss companies Microsoft Copilot is the pragmatic route, because it sits inside Microsoft 365 and the data stays there. The catch: Copilot and ChatGPT do not know your company. Both are the brilliant new employee without a key card.
We build the layer in between that knows your business, and we keep running it. The tool stays yours.
You tell us which work looks the same every week in your company. We tell you whether we can change anything about it and how we would go about it. If we cannot, we say so in the call.
No screen share, no documents, no follow-up email. You speak to Yannick Hirt, not to a sales team.
In the two hours you see the difference on your own task. That is the same day.
The first workflow comes with a date: after four weeks it runs, measured against the figure the two of us took beforehand. In the retainer the further workflows follow over the year.
Three prices are on the site: the Toolbox is free, the two hours cost CHF 750, the retainer starts at CHF 5'000 per month and can be cancelled monthly.
The fourth price, setting up a workflow, comes out of your own figures. Together we work out what the workflow costs you per year today, we say honestly which share of that realistically disappears, and the price is a share of it. Any number we named beforehand would be a guess.
Run it through with your own assumptions, the calculator further up does exactly that.
One named person with a time budget, four hours a week as a guide. That is a condition, not a formality. Without that person every system dies within six months, no matter who built it.
That person does not need to understand AI. They need to know how your business works and to be reachable. For the two hours we need nobody, just ten minutes of preparation from you.
No, and we do not promise you any job cuts either. We measure hours per workflow, not headcount.
The work that stops is the work nobody enjoys. Pulling reports together, typing figures across, giving the same answer for the fifth time. The time goes into work that nobody gets round to today.
The retainer even demands the opposite: one person on your side who is given time for it.
That is why there are two routes. Cloud with a Swiss contract and a data processing agreement, using the models of the large providers, and we tell you before we start which one and where. Or fully in house with open models, without any data leaving the building.
This website runs with an international host. Your company data never runs there.
You do. Readable files, no proprietary format, versioned settings, documented access.
If you cancel, we hand over everything: the company context, the workflows, the profiles, the measurements. You keep all of it and can carry on with somebody else or on your own.
It will change, and faster than you would like. The value sits in the structured knowledge about your business, not in the tool. If the model changes, we migrate. That is part of the retainer and costs nothing extra.
AI-ffective is a Swiss GmbH in formation. Until it is registered, ODCUS AG stands behind it. It is your contracting party, it invoices and it carries the liability. ODCUS AG has been in the commercial register since 2022, registered office in Baar ZG, company number CHE-316.886.794. All of it is in the imprint.
What that means for you: you sign the contract with a registered Swiss stock corporation, not with a statement of intent. The retainer can be cancelled monthly, and everything that gets built belongs to you and stays with you. You are not committing blind for a year.
Rather send a message? Go to the contact form.