UI system designed from the ground up
This user interface went through a rigorous process, from defining default components to designing sector-specific sections; every element was thought through and tested thoroughly. Some elements needed special consideration due to their specific use cases; one such component was a quantity forecast chart for logistics managers at last-mile shipping companies.
Last-mile shipping requires up-to-date and accurate forecasts for the day ahead so logistics managers can book the delivery drivers needed. The primary conflict I had when designing this chart was whether the user should see the algorithms ‘confidence’ level; this is how confident the AI is in its prediction. This could be useful for the user because of how precise the forecast needs to be.
If the logistics manager booked in the specific amount the AI tells them to (the median amount), it could be slightly under what they need - but even one package undelivered is a bad thing! They usually overestimate their needs to make sure all packages are delivered, even if that means some drivers leave empty. This is where the confidence level, or as we ended up calling it, risk level, can help. Saying this, I wanted to make sure the interface remained as simple and usable as possible.
Here are some of the steps I took towards the final version (above). In-between each stage I would discuss the changes with the developers, product managers and UX researchers about where to explore next.