The inspiration for Serene comes from the need for victims of domestic violence to have a discreet and empowering way to call for help.
The article that inspired it all! https://www.cnn.com/2019/11/23/us/911-pizza-domestic-abuse-call/index.html
Serene is a safety application designed to assist women in distress, particularly those experiencing domestic violence or other dangerous situations. It functions by masquerading as a regular e-commerce app, allowing users to maintain privacy and avoid detection by potential abusers. The app features a hidden help button disguised as a notification bell that, when pressed for 3 seconds, activates a discreet alert system. This system enables users to silently contact emergency services or trusted contacts, sending an alert message that includes the user's real-time location and relevant personal information.
"Serene" was meticulously developed through a combination of in-depth product research, AI integration,UI prototyping and strategic use of modern development tools:
Product Research and AI Integration:
Extensive product research guided the design process, focusing on user safety and app discretion. AI and machine learning, via the OpenAI API, were crucial in personalizing the app experience, analyzing user behavior to tailor the UI and features. Development Tools and Technologies:
Flutter: Enabled development across multiple platforms with a single codebase, ensuring a consistent UI. Firebase: We used for real-time data synchronization (Firestore) and secure authentication (Firebase Auth). Dart: The main programming language, key to creating a scalable and robust application. Sinch: Provided essential communication features for emergency alerts. Visual Studio Code: Enhanced the coding process with its comprehensive support for Flutter and Dart. GitHub: Facilitated collaboration and version control, keeping the project's evolution transparent and well-managed.
Challenge for App’s Disguise: How to determine the best disguise for the user and how much agency the user should have over the disguise Solution: We wanted to give the users some input on the type of the disguise but we did not want to overcomplicate the app so we leveraged GenerativeAI by requesting user input on online shopping habits to generate the best disguises for them\
Challenge for Alert Feature: How to disguise the alert feature and make calling for help discreet and easy Solution: We went through a couple different iterations of the alert feature, considering “placing and order” as being a way to alert, but we found this method harder to be fully discreet. We decided we to use a notification button as an intuitive alert button for users to call for help. After conducting user research on this we found that clicking once of the bell is still not discreet enough and there needs to be a decoy (fake pop-up) when you click once and the alert pop up when you click and hold for three seconds (which only the user knows about)
Challenge for AI Images: How to utilise the AI to generate the image each time the user clicked on the refresh button
Solution: This solution required researching APIs that offer real-time, on-demand generative capabilities to ensure seamless integration and responsive performance within the app.