Inspiration
Nowadays every website has at least one chatbot and multiple popups. These make it hard to navigate and can lead to fatigue in the user. Our goal is to combat this using UX design that is more impactful and less intrusive.
What it does
Sixt sense uses personalized recommendations based on the user's and the car's "vibe". These vibes are expressed in a series of tags which will be matched. The user gets assigned a "boring" car in their category. But they get a relatively cheap option to upgrade to a more intersting car. The upgrade option is sold to them using personalized tag lines. If they decline, they will happen to walk past the upgrade again on their route to their assigned car and get the option to upgrade and unlock the car in a single swipe.
How we built it
The app was built using Next.js, the Google Vertex API and GitHub Copilot.
Challenges we ran into
Getting navigation and the compass to work in a browser on a smartphone posed a significant challenge. Also the tagging system required some tuning to give reasonable recommendations.
Accomplishments that we're proud of
The compass and navigation work now and upgrade options get triggered when walking past the cars. The integrations is seamless and provide a pleasant user experience.
What we learned
Using the compass and geolocation api in a browser as well as using AI to find out the "vibe" of a person.
What's next for Sixt Sense
There are multiple ways of improving Sixt Sense. For example images of the upgrade car could be generated on-demand based on the user and where they are going, further personalizing their experience. Upgrades to Protection could be offered to the user together with facts (e.g., about rainy weather, crime rates or just protecting the interior if they have children).
Built With
- copilot
- gemini
- leaflet.js
- next.js
- typescript

Log in or sign up for Devpost to join the conversation.