Taking new projectsWarsaw and Poznań · local timeWe reply within one business day5.0 on Google · 25 reviews
fessburnStart a project

Product design for AI features that people have to trust

People have to trust an AI feature before they act on its answer. We design how the feature tells people they're dealing with AI, and how they can recover when the model gets it wrong.

What we see in ai products products

The problems we fix

The demo isn't the product

A demo is a simulated environment. You choose the input, and you know the expected outcome. On the other hand, real users paste in messy data and take the answer at face value. We test AI flows with realistic inputs, so the failure modes show up before your users find them.

A blank prompt box

So a client we worked with had a content creation tool that opened on a blank prompt box, and new users didn't know what to do to start using it. We designed the onboarding and the conversational guidance, so people could understand what an AI action would do before they used it.

People need to know it's AI

AI Act Article 50 means people have to be told when they're dealing with an AI system. The disclosure is a design problem as much as a legal one: it has to be clear without making people distrust a feature that works.

Rules and deadlines

The rules your product has to meet

  1. Since 2 Aug 2026

    AI Act Article 50 transparency. People have to be told when they're interacting with an AI system.

  2. 2 Dec 2026

    Marking AI-generated content. Generative systems already on the market have to mark the content they produce.

  3. 2 Dec 2027

    High-risk AI systems. Stand-alone high-risk systems take on the AI Act's full obligations, including human oversight.

Start a project

Tell us about your product, and we’ll reply within one business day.

Start a project