Meet Adscan: A smart machine learning tool that predicts how a panel of a hundred people will judge your commercial. Powered by Ster (Dutch broadcaster of radio and television ads).
Challenge: How can we design the Adscan tool in a way that advertisers immediately understand how their commercial is performing and what the advice is.
A UX kick-off and ideation with the team was organised by me. We generated lots of ideas. Ideas that we used to solve this challenge. We constantly balanced between what advertisers want to see and what the smart machine could generate.
We took those insights and I made some first designs in a clickable prototype. We came up with a very simple way of presenting the advice, in a human understandable format: Elements to keep, elements to use less and elements which can be added.
We validated these designs by interviewing different media agencies and incorporated the learnings.
As a result we have a design that humans can understand. Even though it is a machine/ AI that generates the results.