AI agents and automation
Not AI for its own sake: we automate the steps that cost your team the most time, and we say so when it is not worth it.
- Build an appiOS, Android, web
- Software & platformPortal, marketplace, custom software
- AI & automationAgents, workflows, interfaces
- More customersSEO, Google Ads, Meta
- Content & social mediaShoot days, reels, channel care
- WebsiteCompany site, landing pages, shop
- First versionone clear use case
- Extensionextend an existing system
- Large projectseveral parts, many users
- Ongoing supportevery month, no end date
- I don't know yetwe will clarify it in the call
- As soon as possibleit is urgent
- In 1–3 monthsin the coming weeks
- Later this yearplanning ahead
- Just lookingno timeline yet
- Reply within 24 hours
- One contact for everything
- Code and data belong to you
What we automate for you
A task you can name, measured before and after.
AI agents
Answer, sort and prepare — so your team only decides.
Reading documents
Invoices, forms and emails into structured data.
Automated workflows
From enquiry to offer without copying between tools.
Evaluations
Questions to your own data in plain language.
Data protection
Which data goes where is settled in the concept.
Operation
Monitoring and adjustments when your processes change.
What people worry about with AI
Three sentences we hear in almost every first call, and what we answer.
“It makes things up and we do not notice.”
It looks things up and shows its source
An agent that answers customers reads the real account and booking data first. Where there is no source, it says so instead of guessing.
“It will send something embarrassing.”
Nothing goes out unseen
Every reply lands as a draft with a person. You decide later, feature by feature, where the agent may act on its own.
“Our data ends up training someone’s model.”
Business plans, no training
We use the providers’ business plans, where inputs are not used for training, and write down which data leaves your systems at all.
How it works
Four steps from the first call to operation. At each one you know what comes next.
- 1Free · 30 min
First call
Which task costs the most time today.
- 2Free · about 1 week
Concept & offer
What is worth automating, and what is not.
- 34–10 weeks
Build
First on a copy of your data, then in daily use with a person in between.
- 4ongoing
Operation
We watch quality and cost and widen what the agent may do.
Questions about AI projects
Answered briefly. Everything else we clarify in the first call.
Which models do you use?
Whatever fits the task, mostly Claude or GPT, DeepL for translation. We are not tied to a provider, and switching later is planned in from the start.
What does it cost to run?
The models are billed by use. What that means for you, we calculate in the concept, next to the working time it saves.
Does this replace jobs?
In our projects the dull part disappeared: sorting, looking up, copying. The decisions stayed with people, and that is what makes the result usable.
Which task should disappear?
Three questions, then we get back to you with the right people.