Using AI UX patterns

This guide catalogues the UX design patterns applicable to delivering effective AI experiences.

AI has brought with it a new paradigm moving from command to intent based inputs & some of our existing UX patterns are applicable while other completely new ones will be designed.

Real world example are useful, however abstracted examples like these allow us to focus on the core elements & interaction in each pattern, so designers & developers can evaluate & adapt them to their products context.

How LLM's are different

The majority of software utilises the common logic of If > Then statements which are then strung together to achieve complex tasks. Allowing us to code business logic into products for reliable & consistent outcomes.

But...

AI (LLM's) is different, the natural language models aren't hard coded, they're more like a magical black box.

Although extremely powerful, the answers these models provide aren't always accurate. The possibility of hallucinations (lies) mean that the output needs to be reviewed & interpreted by users. This will evolve as models become more reliable, but this is the key difference to consider when crafting AI user experiences.