In what we build for you
We use AI where it does concrete work and ordinary automation where that is enough. Most of what we deliver is rule-driven automation, not models. We tell you when a workflow contains AI, which model is used and what happens if it answers incorrectly.
Anything we build that generates text or content for your customers is labelled as AI-generated, in line with article 50 of the EU AI Act. That applies to chat interfaces too: a user should never have to guess whether they are talking to a person.
With your data
We process your data solely in order to build and test your workflow, never for our own purposes and never to train AI models. We sign a data processing agreement before the engagement starts and tell you exactly which systems the data passes through.
When the engagement is finished we delete or return everything within 30 days.
In what we publish
The texts on this site are written with AI as a tool and reviewed by Andreas Udén, who takes editorial responsibility for every published text. The photographs on the site and the film on the home page are AI-generated.
We do not publish photographs that suggest they document something that never happened, and we do not invent customer quotes, customer logos or colleagues.
What we do not do
- We do not build systems that make decisions about individual people without a person being able to review and change the decision.
- We do not build anything that collects personal data you do not already have the right to process.
- We do not pass your data to third parties beyond the system suppliers required for the workflow to function, and those are listed in the agreement.
If you have questions about any of this, get in touch. We would rather answer an uncomfortable question before an engagement than after it. How we handle data on the site itself is set out in our privacy policy.
This is a translation of the Swedish original. If the two versions differ, the Swedish AI-policy is the one that applies.