What it is about
Whether AI in a product gets used is rarely decided by the model, but by the experience: does the user understand what the AI does? Can they check its result? Do they reach their goal faster than before? Experience design for AI answers these questions before code is written.
What we design
- Guided advice instead of filters - users describe in their own words what they need; the product translates that into models, variants and options.
- Traceable suggestions - every recommendation shows what it is based on and can be accepted, changed or rejected.
- Approvals in the business process - interfaces in which experts review and decide on the work of agents instead of repeating it.
- In your brand’s voice - your brand’s language rules are built in and tested with examples; the AI answers only from approved content and hands topics it should not answer to a person.
- Review before the answer - where a lot of free text is generated, a second model checks every answer against tone and rules before it is shown: release or hold back, with a reason. Held-back answers are logged and flagged together with the question, so your team can refine in a targeted way.
- Relevance over completeness - content that fits the user’s situation instead of showing everything at once.
- Design systems - components that work the same in website, app and assistant.
Where we have done it
Product configurators and customisation, a digital owner’s manual, guided product advice and launch experiences - the examples are under References.
For whom
For companies whose products and services come with many variants, options or questions - from machinery and insurance to medical technology - and who want to use AI in a way that customers and employees actually use.
